code generator and the set of reusable components that can be used to build
target-specific backends. The two most important interfaces (<a
href="#targetmachine"><tt>TargetMachine</tt></a> and <a
-href="#targetdata"><tt>TargetData</tt></a> classes) are the only ones that are
+href="#targetdata"><tt>TargetData</tt></a>) are the only ones that are
required to be defined for a backend to fit into the LLVM system, but the others
must be defined if the reusable code generator components are going to be
used.</p>
represent any required register assignments due to target constraints or calling
conventions.</li>
-<li><b>SSA-based Machine Code Optimizations</b> - This (optional) stage consists
-of a series of machine-code optimizations that operate on the SSA-form produced
-by the instruction selector. Optimizations like modulo-scheduling, normal
-scheduling, or peephole optimization work here.</li>
+<li><b><a href="#ssamco">SSA-based Machine Code Optimizations</a></b> - This
+optional stage consists of a series of machine-code optimizations that
+operate on the SSA-form produced by the instruction selector. Optimizations
+like modulo-scheduling, normal scheduling, or peephole optimization work here.
+</li>
-<li><b>Register Allocation</b> - The target code is transformed from an infinite
-virtual register file in SSA form to the concrete register file used by the
-target. This phase introduces spill code and eliminates all virtual register
-references from the program.</li>
+<li><b><a name="#regalloc">Register Allocation</a></b> - The
+target code is transformed from an infinite virtual register file in SSA form
+to the concrete register file used by the target. This phase introduces spill
+code and eliminates all virtual register references from the program.</li>
-<li><b>Prolog/Epilog Code Insertion</b> - Once the machine code has been
-generated for the function and the amount of stack space required is known (used
-for LLVM alloca's and spill slots), the prolog and epilog code for the function
-can be inserted and "abstract stack location references" can be eliminated.
-This stage is responsible for implementing optimizations like frame-pointer
-elimination and stack packing.</li>
+<li><b><a name="#proepicode">Prolog/Epilog Code Insertion</a></b> - Once the
+machine code has been generated for the function and the amount of stack space
+required is known (used for LLVM alloca's and spill slots), the prolog and
+epilog code for the function can be inserted and "abstract stack location
+references" can be eliminated. This stage is responsible for implementing
+optimizations like frame-pointer elimination and stack packing.</li>
-<li><b>Late Machine Code Optimizations</b> - Optimizations that operate on
-"final" machine code can go here, such as spill code scheduling and peephole
-optimizations.</li>
+<li><b><a name="latemco">Late Machine Code Optimizations</a></b> - Optimizations
+that operate on "final" machine code can go here, such as spill code scheduling
+and peephole optimizations.</li>
-<li><b>Code Emission</b> - The final stage actually outputs the code for
-the current function, either in the target assembler format or in machine
-code.</li>
+<li><b><a name="codemission">Code Emission</a></b> - The final stage actually
+puts out the code for the current function, either in the target assembler
+format or in machine code.</li>
</ol>
<p>The target description classes require a detailed description of the target
architecture. These target descriptions often have a large amount of common
-information (e.g., an add instruction is almost identical to a sub instruction).
+information (e.g., an <tt>add</tt> instruction is almost identical to a
+<tt>sub</tt> instruction).
In order to allow the maximum amount of commonality to be factored out, the LLVM
code generator uses the <a href="TableGenFundamentals.html">TableGen</a> tool to
-describe big chunks of the target machine, which allows the use of domain- and
-target-specific abstractions to reduce the amount of repetition.
+describe big chunks of the target machine, which allows the use of
+domain-specific and target-specific abstractions to reduce the amount of
+repetition.
</p>
</div>
<p>The LLVM target description classes (which are located in the
<tt>include/llvm/Target</tt> directory) provide an abstract description of the
-target machine, independent of any particular client. These classes are
-designed to capture the <i>abstract</i> properties of the target (such as what
-instruction and registers it has), and do not incorporate any particular pieces
-of code generation algorithms (these interfaces do not take interference graphs
-as inputs or other algorithm-specific data structures).</p>
+target machine; independent of any particular client. These classes are
+designed to capture the <i>abstract</i> properties of the target (such as the
+instructions and registers it has), and do not incorporate any particular pieces
+of code generation algorithms. These interfaces do not take interference graphs
+as inputs or other algorithm-specific data structures.</p>
<p>All of the target description classes (except the <tt><a
href="#targetdata">TargetData</a></tt> class) are designed to be subclassed by
<p>The <tt>TargetMachine</tt> class provides virtual methods that are used to
access the target-specific implementations of the various target description
-classes (with the <tt>getInstrInfo</tt>, <tt>getRegisterInfo</tt>,
-<tt>getFrameInfo</tt>, ... methods). This class is designed to be specialized by
+classes via the <tt>get*Info</tt> methods (<tt>getInstrInfo</tt>,
+<tt>getRegisterInfo</tt>, <tt>getFrameInfo</tt>, etc.). This class is
+designed to be specialized by
a concrete target implementation (e.g., <tt>X86TargetMachine</tt>) which
implements the various virtual methods. The only required target description
class is the <a href="#targetdata"><tt>TargetData</tt></a> class, but if the
<div class="doc_text">
<p>The <tt>TargetData</tt> class is the only required target description class,
-and it is the only class that is not extensible (it cannot be derived from). It
-specifies information about how the target lays out memory for structures, the
-alignment requirements for various data types, the size of pointers in the
-target, and whether the target is little- or big-endian.</p>
+and it is the only class that is not extensible. You cannot derived a new
+class from it. <tt>TargetData</tt> specifies information about how the target
+lays out memory for structures, the alignment requirements for various data
+types, the size of pointers in the target, and whether the target is
+little-endian or big-endian.</p>
</div>
<p>The <tt>TargetLowering</tt> class is used by SelectionDAG based instruction
selectors primarily to describe how LLVM code should be lowered to SelectionDAG
-operations. Among other things, this class indicates an initial register class
-to use for various ValueTypes, which operations are natively supported by the
-target machine, and some other miscellaneous properties (such as the return type
-of setcc operations, the type to use for shift amounts, etc).</p>
+operations. Among other things, this class indicates:
+<ul><li>an initial register class to use for various ValueTypes,</li>
+ <li>which operations are natively supported by the target machine,</li>
+ <li>the return type of setcc operations, and</li>
+ <li>the type to use for shift amounts, etc</li>.
+</ol></p>
</div>
<p>Target machine instructions are represented as instances of the
<tt>MachineInstr</tt> class. This class is an extremely abstract way of
-representing machine instructions. In particular, all it keeps track of is
-an opcode number and some number of operands.</p>
+representing machine instructions. In particular, it only keeps track of
+an opcode number and a set of operands.</p>
-<p>The opcode number is an simple unsigned number that only has meaning to a
+<p>The opcode number is a simple unsigned number that only has meaning to a
specific backend. All of the instructions for a target should be defined in
-the <tt>*InstrInfo.td</tt> file for the target, and the opcode enum values
+the <tt>*InstrInfo.td</tt> file for the target. The opcode enum values
are auto-generated from this description. The <tt>MachineInstr</tt> class does
not have any information about how to interpret the instruction (i.e., what the
semantics of the instruction are): for that you must refer to the
the operands should be stored as "<tt>%i3, %i1, %i2</tt>": with the destination
first.</p>
-<p>Keeping destination operands at the beginning of the operand list has several
-advantages. In particular, the debugging printer will print the instruction
-like this:</p>
+<p>Keeping destination (definition) operands at the beginning of the operand
+list has several advantages. In particular, the debugging printer will print
+the instruction like this:</p>
<pre>
%r3 = add %i1, %i2
<p>
The key thing to remember with the <tt>BuildMI</tt> functions is that you have
-to specify the number of operands that the machine instruction will take
-(allowing efficient memory allocation). Also, if operands default to be uses
-of values, not definitions. If you need to add a definition operand (other
-than the optional destination register), you must explicitly mark it as such.
+to specify the number of operands that the machine instruction will take. This
+allows for efficient memory allocation. You also need to specify if operands
+default to be uses of values, not definitions. If you need to add a definition
+operand (other than the optional destination register), you must explicitly
+mark it as such.
</p>
</div>
<!-- _______________________________________________________________________ -->
<div class="doc_subsubsection">
- <a name="fixedregs">Fixed (aka preassigned) registers</a>
+ <a name="fixedregs">Fixed (preassigned) registers</a>
</div>
<div class="doc_text">
presence of fixed registers. In particular, there are often places in the
instruction stream where the register allocator <em>must</em> arrange for a
particular value to be in a particular register. This can occur due to
-limitations in the instruction set (e.g., the X86 can only do a 32-bit divide
+limitations of the instruction set (e.g., the X86 can only do a 32-bit divide
with the <tt>EAX</tt>/<tt>EDX</tt> registers), or external factors like calling
conventions. In any case, the instruction selector should emit code that
copies a virtual register into or out of a physical register when needed.</p>
<div class="doc_text">
-<p><tt>MachineInstr</tt>'s are initially instruction selected in SSA-form, and
+<p><tt>MachineInstr</tt>'s are initially selected in SSA-form, and
are maintained in SSA-form until register allocation happens. For the most
part, this is trivially simple since LLVM is already in SSA form: LLVM PHI nodes
become machine code PHI nodes, and virtual registers are only allowed to have a
<div class="doc_text">
<p>
-Instruction Selection is the process of translating the LLVM code input to the
+Instruction Selection is the process of translating LLVM code presented to the
code generator into target-specific machine instructions. There are several
well-known ways to do this in the literature. In LLVM there are two main forms:
the old-style 'simple' instruction selector (which effectively peephole selects
SelectionDAG infrastructure.</p>
<p>In time, most of the target-specific code for instruction selection will be
-auto-generated from the target .td files. For now, however, the <a
-href="#selectiondag_select">Select Phase</a> must still be written by hand.</p>
+auto-generated from the target description (<tt>*.td</tt>) files. For now,
+however, the <a href="#selectiondag_select">Select Phase</a> must still be
+written by hand.</p>
</div>
<!-- _______________________________________________________________________ -->
<div class="doc_text">
<p>
-The SelectionDAG provides an abstraction for representing code in a way that is
-amenable to instruction selection using automatic techniques
-(e.g. dynamic-programming based optimal pattern matching selectors), as well as
-an abstraction that is useful for other phases of code generation (in
-particular, instruction scheduling). Additionally, the SelectionDAG provides a
-host representation where a large variety of very-low-level (but
-target-independent) <a href="#selectiondag_optimize">optimizations</a> may be
+The SelectionDAG provides an abstraction for code representation in a way that
+is amenable to instruction selection using automatic techniques
+(e.g. dynamic-programming based optimal pattern matching selectors), It is also
+well suited to other phases of code generation; in particular, instruction scheduling. Additionally, the SelectionDAG provides a host representation where a
+large variety of very-low-level (but target-independent)
+<a href="#selectiondag_optimize">optimizations</a> may be
performed: ones which require extensive information about the instructions
efficiently supported by the target.
</p>
<p>
The SelectionDAG is a Directed-Acyclic-Graph whose nodes are instances of the
-<tt>SDNode</tt> class. The primary payload of the Node is its operation code
-(Opcode) that indicates what the operation the node performs. The various
-operation node types are described at the top of the
-<tt>include/llvm/CodeGen/SelectionDAGNodes.h</tt> file. Depending on the operation, nodes may contain additional information (e.g. the condition code
+<tt>SDNode</tt> class. The primary payload of the <tt>SDNode</tt> is its
+operation code (Opcode) that indicates what operation the node performs.
+The various operation node types are described at the top of the
+<tt>include/llvm/CodeGen/SelectionDAGNodes.h</tt> file. Depending on the
+operation, nodes may contain additional information (e.g. the condition code
for a SETCC node) contained in a derived class.</p>
-<p>Each node in the graph may define multiple values
-(e.g. for a combined div/rem operation and many other situations), though most
-operations define a single value. Each node also has some number of operands,
-which are edges to the node defining the used value. Because nodes may define
-multiple values, edges are represented by instances of the <tt>SDOperand</tt>
-class, which is a <SDNode, unsigned> pair, indicating the node and result
-value being used. Each value produced by a SDNode has an associated
-MVT::ValueType, indicating what type the value is.
+<p>Although most operations define a single value, each node in the graph may
+define multiple values. For example, a combined div/rem operation will define
+both the dividend and the remainder. Many other situations require multiple
+values as well. Each node also has some number of operands, which are edges
+to the node defining the used value. Because nodes may define multiple values,
+edges are represented by instances of the <tt>SDOperand</tt> class, which is
+a <SDNode, unsigned> pair, indicating the node and result
+value being used, respectively. Each value produced by an SDNode has an
+associated MVT::ValueType, indicating what type the value is.
</p>
<p>
-SelectionDAGs contain two different kinds of value: those that represent data
+SelectionDAGs contain two different kinds of values: those that represent data
flow and those that represent control flow dependencies. Data values are simple
-edges with a integer or floating point value type. Control edges are
+edges with an integer or floating point value type. Control edges are
represented as "chain" edges which are of type MVT::Other. These edges provide
an ordering between nodes that have side effects (such as
loads/stores/calls/return/etc). All nodes that have side effects should take a
<p>
A SelectionDAG has designated "Entry" and "Root" nodes. The Entry node is
-always a marker node with Opcode of ISD::TokenFactor. The Root node is the
-final side effecting node in the token chain (for example, in a single basic
-block function, this would be the return node).
+always a marker node with an Opcode of ISD::TokenFactor. The Root node is the
+final side-effecting node in the token chain. For example, in a single basic
+block function, this would be the return node.
</p>
<p>
-One important concept for SelectionDAGs is the notion of a "legal" vs "illegal"
+One important concept for SelectionDAGs is the notion of a "legal" vs. "illegal"
DAG. A legal DAG for a target is one that only uses supported operations and
supported types. On PowerPC, for example, a DAG with any values of i1, i8, i16,
or i64 type would be illegal. The <a href="#selectiondag_legalize">legalize</a>
-phase is the one responsible for turning an illegal DAG into a legal DAG.
+phase is responsible for turning an illegal DAG into a legal DAG.
</p>
</div>
<!-- _______________________________________________________________________ -->
<div class="doc_subsubsection">
- <a name="selectiondag_process">SelectionDAG Code Generation Process</a>
+ <a name="selectiondag_process">SelectionDAG Instruction Selection Process</a>
</div>
<div class="doc_text">
performs simple optimizations on the SelectionDAG to simplify it and
recognize meta instructions (like rotates and div/rem pairs) for
targets that support these meta operations. This makes the resultant code
- more efficient and the 'select' phase more simple.
+ more efficient and the 'select instructions from DAG' phase (below) simpler.
</li>
<li><a href="#selectiondag_legalize">Legalize SelectionDAG</a> - This stage
converts the illegal SelectionDAG to a legal SelectionDAG, by eliminating
<p>
The initial SelectionDAG is naively peephole expanded from the LLVM input by
-the SelectionDAGLowering class in the SelectionDAGISel.cpp file. The idea of
-doing this pass is to expose as much low-level target-specific details to the
-SelectionDAG as possible. This pass is mostly hard-coded (e.g. an LLVM add
-turns into a SDNode add, a geteelementptr is expanded into the obvious
-arithmetic, etc) but does require target-specific hooks to lower calls and
+the <tt>SelectionDAGLowering</tt> class in the SelectionDAGISel.cpp file. The
+intent of this pass is to expose as much low-level, target-specific details
+to the SelectionDAG as possible. This pass is mostly hard-coded (e.g. an LLVM
+add turns into an SDNode add while a geteelementptr is expanded into the obvious
+arithmetic). This pass requires target-specific hooks to lower calls and
returns, varargs, etc. For these features, the TargetLowering interface is
used.
</p>
<ol>
<li><p>Convert values of unsupported types to values of supported types.</p>
<p>There are two main ways of doing this: promoting a small type to a larger
- type (e.g. f32 -> f64, or i16 -> i32), and expanding larger integer types
+ type (e.g. f32 -> f64, or i16 -> i32), and breaking up large
+ integer types
to smaller ones (e.g. implementing i64 with i32 operations where
- possible). Promotion insert sign and zero extensions as needed to make
- sure that the final code has the same behavior as the input.</p>
+ possible). Type conversions can insert sign and zero extensions as
+ needed to make sure that the final code has the same behavior as the
+ input.</p>
</li>
<li><p>Eliminate operations that are not supported by the target in a supported
conditional moves). Legalize takes care of either open-coding another
sequence of operations to emulate the operation (this is known as
expansion), promoting to a larger type that supports the operation
- (promotion), or can use a target-specific hook to implement the
+ (promotion), or using a target-specific hook to implement the
legalization.</p>
</li>
</ol>
<p>
Instead of using a Legalize pass, we could require that every target-specific
-<a href="#selectiondag_optimize">selector</a> support and expand every operator
-and type even if they are not supported and may require many instructions to
-implement (in fact, this is the approach taken by the "simple" selectors).
-However, using a Legalize pass allows all of the cannonicalization patterns to
-be shared across targets, and makes it very easy to optimize the cannonicalized
-code (because it is still in the form of a DAG).
+<a href="#selectiondag_optimize">selector</a> supports and expands every
+operator and type even if they are not supported and may require many
+instructions to implement (in fact, this is the approach taken by the
+"simple" selectors). However, using a Legalize pass allows all of the
+cannonicalization patterns to be shared across targets which makes it very
+easy to optimize the cannonicalized code because it is still in the form of
+a DAG.
</p>
</div>
<p>
The SelectionDAG optimization phase is run twice for code generation: once
-immediately after the DAG is built and once after legalization. The first pass
-allows the initial code to be cleaned up, (for example) performing optimizations
-that depend on knowing that the operators have restricted type inputs. The second
-pass cleans up the messy code generated by the Legalize pass, allowing Legalize to
-be very simple (not having to take into account many special cases.
+immediately after the DAG is built and once after legalization. The first run
+of the pass allows the initial code to be cleaned up (e.g. performing
+optimizations that depend on knowing that the operators have restricted type
+inputs). The second run of the pass cleans up the messy code generated by the
+Legalize pass, allowing Legalize to be very simple since it can ignore many
+special cases.
</p>
<p>
<p>The Select phase is the bulk of the target-specific code for instruction
selection. This phase takes a legal SelectionDAG as input, and does simple
pattern matching on the DAG to generate code. In time, the Select phase will
-be automatically generated from the targets InstrInfo.td file, which is why we
-want to make the Select phase a simple and mechanical as possible.</p>
+be automatically generated from the target's InstrInfo.td file, which is why we
+want to make the Select phase as simple and mechanical as possible.</p>
</div>
<ol>
<li>Optional whole-function selection.</li>
<li>Select is a graph translation phase.</li>
-<li>Place the machine instrs resulting from Select according to register pressure or a schedule.</li>
+<li>Place the machine instructions resulting from Select according to register
+pressure or a schedule.</li>
<li>DAG Scheduling.</li>
-<li>Auto-generate the Select phase from the target .td files.</li>
+<li>Auto-generate the Select phase from the target description (*.td) files.
+</li>
</ol>
</div>
-
+
+<!-- ======================================================================= -->
+<div class="doc_subsection">
+ <a name="ssamco">SSA-based Machine Code Optimizations</a>
+</div>
+<div class="doc_text"><p>To Be Written</p></div>
+<!-- ======================================================================= -->
+<div class="doc_subsection">
+ <a name="regalloc">Register Allocation</a>
+</div>
+<div class="doc_text"><p>To Be Written</p></div>
+<!-- ======================================================================= -->
+<div class="doc_subsection">
+ <a name="proepicode">Prolog/Epilog Code Insertion</a>
+</div>
+<div class="doc_text"><p>To Be Written</p></div>
+<!-- ======================================================================= -->
+<div class="doc_subsection">
+ <a name="latemco">Late Machine Code Optimizations</a>
+</div>
+<div class="doc_text"><p>To Be Written</p></div>
+<!-- ======================================================================= -->
+<div class="doc_subsection">
+ <a name="codemission">Code Emission</a>
+</div>
<!-- *********************************************************************** -->
<div class="doc_section">
<div class="doc_text">
-<p>This section of the document explains any features or design decisions that
+<p>This section of the document explains features or design decisions that
are specific to the code generator for a particular target.</p>
</div>
<div class="doc_text">
-<p>
-The x86 has a very, uhm, flexible, way of accessing memory. It is capable of
+<p>The x86 has a very flexible way of accessing memory. It is capable of
forming memory addresses of the following expression directly in integer
instructions (which use ModR/M addressing):</p>
Base+[1,2,4,8]*IndexReg+Disp32
</pre>
-<p>Wow, that's crazy. In order to represent this, LLVM tracks no less than 4
-operands for each memory operand of this form. This means that the "load" form
-of 'mov' has the following "Operands" in this order:</p>
+<p>In order to represent this, LLVM tracks no less than 4 operands for each
+memory operand of this form. This means that the "load" form of 'mov' has the
+following <tt>MachineOperand</tt>s in this order:</p>
<pre>
Index: 0 | 1 2 3 4
OperandTy: VirtReg, | VirtReg, UnsImm, VirtReg, SignExtImm
</pre>
-<p>Stores and all other instructions treat the four memory operands in the same
-way, in the same order.</p>
+<p>Stores, and all other instructions, treat the four memory operands in the
+same way, in the same order.</p>
</div>
<div class="doc_text">
<p>
-An instruction name consists of the base name, a default operand size
-followed by a character per operand with an optional special size. For
-example:</p>
+An instruction name consists of the base name, a default operand size, and a
+a character per operand with an optional special size. For example:</p>
<p>
<tt>ADD8rr</tt> -> add, 8-bit register, 8-bit register<br>