huangmeng
/
Embedded_AI


			
							123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297
							<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<!-- Copyright (C) 1988-2017 Free Software Foundation, Inc.

Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with the
Invariant Sections being "Funding Free Software", the Front-Cover
Texts being (a) (see below), and with the Back-Cover Texts being (b)
(see below).  A copy of the license is included in the section entitled
"GNU Free Documentation License".

(a) The FSF's Front-Cover Text is:

A GNU Manual

(b) The FSF's Back-Cover Text is:

You have freedom to copy and modify this GNU Manual, like GNU
     software.  Copies published by the Free Software Foundation raise
     funds for GNU development. -->
<!-- Created by GNU Texinfo 5.2, http://www.gnu.org/software/texinfo/ -->
<head>
<title>GNU Compiler Collection (GCC) Internals: Tuple representation</title>

<meta name="description" content="GNU Compiler Collection (GCC) Internals: Tuple representation">
<meta name="keywords" content="GNU Compiler Collection (GCC) Internals: Tuple representation">
<meta name="resource-type" content="document">
<meta name="distribution" content="global">
<meta name="Generator" content="makeinfo">
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<link href="index.html#Top" rel="start" title="Top">
<link href="Option-Index.html#Option-Index" rel="index" title="Option Index">
<link href="index.html#SEC_Contents" rel="contents" title="Table of Contents">
<link href="GIMPLE.html#GIMPLE" rel="up" title="GIMPLE">
<link href="Class-hierarchy-of-GIMPLE-statements.html#Class-hierarchy-of-GIMPLE-statements" rel="next" title="Class hierarchy of GIMPLE statements">
<link href="GIMPLE.html#GIMPLE" rel="prev" title="GIMPLE">
<style type="text/css">
<!--
a.summary-letter {text-decoration: none}
blockquote.smallquotation {font-size: smaller}
div.display {margin-left: 3.2em}
div.example {margin-left: 3.2em}
div.indentedblock {margin-left: 3.2em}
div.lisp {margin-left: 3.2em}
div.smalldisplay {margin-left: 3.2em}
div.smallexample {margin-left: 3.2em}
div.smallindentedblock {margin-left: 3.2em; font-size: smaller}
div.smalllisp {margin-left: 3.2em}
kbd {font-style:oblique}
pre.display {font-family: inherit}
pre.format {font-family: inherit}
pre.menu-comment {font-family: serif}
pre.menu-preformatted {font-family: serif}
pre.smalldisplay {font-family: inherit; font-size: smaller}
pre.smallexample {font-size: smaller}
pre.smallformat {font-family: inherit; font-size: smaller}
pre.smalllisp {font-size: smaller}
span.nocodebreak {white-space:nowrap}
span.nolinebreak {white-space:nowrap}
span.roman {font-family:serif; font-weight:normal}
span.sansserif {font-family:sans-serif; font-weight:normal}
ul.no-bullet {list-style: none}
-->
</style>


</head>

<body lang="en" bgcolor="#FFFFFF" text="#000000" link="#0000FF" vlink="#800080" alink="#FF0000">
<a name="Tuple-representation"></a>
<div class="header">
<p>
Next: <a href="Class-hierarchy-of-GIMPLE-statements.html#Class-hierarchy-of-GIMPLE-statements" accesskey="n" rel="next">Class hierarchy of GIMPLE statements</a>, Up: <a href="GIMPLE.html#GIMPLE" accesskey="u" rel="up">GIMPLE</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Option-Index.html#Option-Index" title="Index" rel="index">Index</a>]</p>
</div>
<hr>
<a name="Tuple-representation-1"></a>
<h3 class="section">11.1 Tuple representation</h3>
<a name="index-tuples"></a>

<p>GIMPLE instructions are tuples of variable size divided in two
groups: a header describing the instruction and its locations,
and a variable length body with all the operands. Tuples are
organized into a hierarchy with 3 main classes of tuples.
</p>
<a name="gimple-_0028gsbase_0029"></a>
<h4 class="subsection">11.1.1 <code>gimple</code> (gsbase)</h4>
<a name="index-gimple"></a>

<p>This is the root of the hierarchy, it holds basic information
needed by most GIMPLE statements. There are some fields that
may not be relevant to every GIMPLE statement, but those were
moved into the base structure to take advantage of holes left by
other fields (thus making the structure more compact).  The
structure takes 4 words (32 bytes) on 64 bit hosts:
</p>
<table>
<tr><td>Field</td><td>Size (bits)</td></tr>
<tr><td><code>code</code></td><td>8</td></tr>
<tr><td><code>subcode</code></td><td>16</td></tr>
<tr><td><code>no_warning</code></td><td>1</td></tr>
<tr><td><code>visited</code></td><td>1</td></tr>
<tr><td><code>nontemporal_move</code></td><td>1</td></tr>
<tr><td><code>plf</code></td><td>2</td></tr>
<tr><td><code>modified</code></td><td>1</td></tr>
<tr><td><code>has_volatile_ops</code></td><td>1</td></tr>
<tr><td><code>references_memory_p</code></td><td>1</td></tr>
<tr><td><code>uid</code></td><td>32</td></tr>
<tr><td><code>location</code></td><td>32</td></tr>
<tr><td><code>num_ops</code></td><td>32</td></tr>
<tr><td><code>bb</code></td><td>64</td></tr>
<tr><td><code>block</code></td><td>63</td></tr>
<tr><td>Total size</td><td>32 bytes</td></tr>
</table>

<ul>
<li> <code>code</code>
Main identifier for a GIMPLE instruction.

</li><li> <code>subcode</code>
Used to distinguish different variants of the same basic
instruction or provide flags applicable to a given code. The
<code>subcode</code> flags field has different uses depending on the code of
the instruction, but mostly it distinguishes instructions of the
same family. The most prominent use of this field is in
assignments, where subcode indicates the operation done on the
RHS of the assignment. For example, a = b + c is encoded as
<code>GIMPLE_ASSIGN &lt;PLUS_EXPR, a, b, c&gt;</code>.

</li><li> <code>no_warning</code>
Bitflag to indicate whether a warning has already been issued on
this statement.

</li><li> <code>visited</code>
General purpose &ldquo;visited&rdquo; marker. Set and cleared by each pass
when needed.

</li><li> <code>nontemporal_move</code>
Bitflag used in assignments that represent non-temporal moves.
Although this bitflag is only used in assignments, it was moved
into the base to take advantage of the bit holes left by the
previous fields.

</li><li> <code>plf</code>
Pass Local Flags. This 2-bit mask can be used as general purpose
markers by any pass. Passes are responsible for clearing and
setting these two flags accordingly.

</li><li> <code>modified</code>
Bitflag to indicate whether the statement has been modified.
Used mainly by the operand scanner to determine when to re-scan a
statement for operands.

</li><li> <code>has_volatile_ops</code>
Bitflag to indicate whether this statement contains operands that
have been marked volatile.

</li><li> <code>references_memory_p</code>
Bitflag to indicate whether this statement contains memory
references (i.e., its operands are either global variables, or
pointer dereferences or anything that must reside in memory).

</li><li> <code>uid</code>
This is an unsigned integer used by passes that want to assign
IDs to every statement. These IDs must be assigned and used by
each pass.

</li><li> <code>location</code>
This is a <code>location_t</code> identifier to specify source code
location for this statement. It is inherited from the front
end.

</li><li> <code>num_ops</code>
Number of operands that this statement has. This specifies the
size of the operand vector embedded in the tuple. Only used in
some tuples, but it is declared in the base tuple to take
advantage of the 32-bit hole left by the previous fields.

</li><li> <code>bb</code>
Basic block holding the instruction.

</li><li> <code>block</code>
Lexical block holding this statement.  Also used for debug
information generation.
</li></ul>

<a name="gimple_005fstatement_005fwith_005fops"></a>
<h4 class="subsection">11.1.2 <code>gimple_statement_with_ops</code></h4>
<a name="index-gimple_005fstatement_005fwith_005fops"></a>

<p>This tuple is actually split in two:
<code>gimple_statement_with_ops_base</code> and
<code>gimple_statement_with_ops</code>. This is needed to accommodate the
way the operand vector is allocated. The operand vector is
defined to be an array of 1 element. So, to allocate a dynamic
number of operands, the memory allocator (<code>gimple_alloc</code>) simply
allocates enough memory to hold the structure itself plus <code>N
- 1</code> operands which run &ldquo;off the end&rdquo; of the structure. For
example, to allocate space for a tuple with 3 operands,
<code>gimple_alloc</code> reserves <code>sizeof (struct
gimple_statement_with_ops) + 2 * sizeof (tree)</code> bytes.
</p>
<p>On the other hand, several fields in this tuple need to be shared
with the <code>gimple_statement_with_memory_ops</code> tuple. So, these
common fields are placed in <code>gimple_statement_with_ops_base</code> which
is then inherited from the other two tuples.
</p>

<table>
<tr><td><code>gsbase</code></td><td>256</td></tr>
<tr><td><code>def_ops</code></td><td>64</td></tr>
<tr><td><code>use_ops</code></td><td>64</td></tr>
<tr><td><code>op</code></td><td><code>num_ops</code> * 64</td></tr>
<tr><td>Total size</td><td>48 + 8 * <code>num_ops</code> bytes</td></tr>
</table>

<ul>
<li> <code>gsbase</code>
Inherited from <code>struct gimple</code>.

</li><li> <code>def_ops</code>
Array of pointers into the operand array indicating all the slots that
contain a variable written-to by the statement. This array is
also used for immediate use chaining. Note that it would be
possible to not rely on this array, but the changes required to
implement this are pretty invasive.

</li><li> <code>use_ops</code>
Similar to <code>def_ops</code> but for variables read by the statement.

</li><li> <code>op</code>
Array of trees with <code>num_ops</code> slots.
</li></ul>

<a name="gimple_005fstatement_005fwith_005fmemory_005fops"></a>
<h4 class="subsection">11.1.3 <code>gimple_statement_with_memory_ops</code></h4>

<p>This tuple is essentially identical to <code>gimple_statement_with_ops</code>,
except that it contains 4 additional fields to hold vectors
related memory stores and loads.  Similar to the previous case,
the structure is split in two to accommodate for the operand
vector (<code>gimple_statement_with_memory_ops_base</code> and
<code>gimple_statement_with_memory_ops</code>).
</p>

<table>
<tr><td>Field</td><td>Size (bits)</td></tr>
<tr><td><code>gsbase</code></td><td>256</td></tr>
<tr><td><code>def_ops</code></td><td>64</td></tr>
<tr><td><code>use_ops</code></td><td>64</td></tr>
<tr><td><code>vdef_ops</code></td><td>64</td></tr>
<tr><td><code>vuse_ops</code></td><td>64</td></tr>
<tr><td><code>stores</code></td><td>64</td></tr>
<tr><td><code>loads</code></td><td>64</td></tr>
<tr><td><code>op</code></td><td><code>num_ops</code> * 64</td></tr>
<tr><td>Total size</td><td>80 + 8 * <code>num_ops</code> bytes</td></tr>
</table>

<ul>
<li> <code>vdef_ops</code>
Similar to <code>def_ops</code> but for <code>VDEF</code> operators. There is
one entry per memory symbol written by this statement. This is
used to maintain the memory SSA use-def and def-def chains.

</li><li> <code>vuse_ops</code>
Similar to <code>use_ops</code> but for <code>VUSE</code> operators. There is
one entry per memory symbol loaded by this statement. This is
used to maintain the memory SSA use-def chains.

</li><li> <code>stores</code>
Bitset with all the UIDs for the symbols written-to by the
statement.  This is different than <code>vdef_ops</code> in that all the
affected symbols are mentioned in this set.  If memory
partitioning is enabled, the <code>vdef_ops</code> vector will refer to memory
partitions. Furthermore, no SSA information is stored in this
set.

</li><li> <code>loads</code>
Similar to <code>stores</code>, but for memory loads. (Note that there
is some amount of redundancy here, it should be possible to
reduce memory utilization further by removing these sets).
</li></ul>

<p>All the other tuples are defined in terms of these three basic
ones. Each tuple will add some fields.
</p>

<hr>
<div class="header">
<p>
Next: <a href="Class-hierarchy-of-GIMPLE-statements.html#Class-hierarchy-of-GIMPLE-statements" accesskey="n" rel="next">Class hierarchy of GIMPLE statements</a>, Up: <a href="GIMPLE.html#GIMPLE" accesskey="u" rel="up">GIMPLE</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Option-Index.html#Option-Index" title="Index" rel="index">Index</a>]</p>
</div>


</body>
</html>