<!DOCTYPE html>
<html lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<meta name="generator" content="AsciiDoc 8.6.9">
<title>ToMachine</title>
<link rel="stylesheet" href="./asciidoc.css" type="text/css">
<link rel="stylesheet" href="./pygments.css" type="text/css">


<script type="text/javascript" src="./asciidoc.js"></script>
<script type="text/javascript">
/*<![CDATA[*/
asciidoc.install();
/*]]>*/
</script>
<link rel="stylesheet" href="./mlton.css" type="text/css">
</head>
<body class="article">
<div id="banner">
<div id="banner-home">
<a href="./Home">MLton 20180207</a>
</div>
</div>
<div id="header">
<h1>ToMachine</h1>
</div>
<div id="content">
<div id="preamble">
<div class="sectionbody">
<div class="paragraph"><p><a href="ToMachine">ToMachine</a> is a translation pass from the <a href="RSSA">RSSA</a>
<a href="IntermediateLanguage">IntermediateLanguage</a> to the <a href="Machine">Machine</a> <a href="IntermediateLanguage">IntermediateLanguage</a>.</p></div>
</div>
</div>
<div class="sect1">
<h2 id="_description">Description</h2>
<div class="sectionbody">
<div class="paragraph"><p>This pass converts from a <a href="RSSA">RSSA</a> program into a <a href="Machine">Machine</a> program.</p></div>
<div class="paragraph"><p>It uses <a href="AllocateRegisters">AllocateRegisters</a>, <a href="Chunkify">Chunkify</a>, and <a href="ParallelMove">ParallelMove</a>.</p></div>
</div>
</div>
<div class="sect1">
<h2 id="_implementation">Implementation</h2>
<div class="sectionbody">
<div class="ulist"><ul>
<li>
<p>
<a href="https://github.com/MLton/mlton/blob/master/mlton/backend/backend.sig"><span class="monospaced">backend.sig</span></a>
</p>
</li>
<li>
<p>
<a href="https://github.com/MLton/mlton/blob/master/mlton/backend/backend.fun"><span class="monospaced">backend.fun</span></a>
</p>
</li>
</ul></div>
</div>
</div>
<div class="sect1">
<h2 id="_details_and_notes">Details and Notes</h2>
<div class="sectionbody">
<div class="paragraph"><p>Because the MLton runtime system is shared by all <a href="Codegen">codegens</a>, it is most
convenient to decide on stack layout <em>before</em> any <a href="Codegen">codegen</a> takes over.
In particular, we compute all the stack frame info for each <a href="RSSA">RSSA</a>
function, including stack size, <a href="GarbageCollection">garbage collector</a>
masks for each frame, etc.  To do so, the <a href="Machine">Machine</a>
<a href="IntermediateLanguage">IntermediateLanguage</a> imagines an abstract machine with an infinite
number of (pseudo-)registers of every size.  A liveness analysis
determines, for each variable, whether or not it is live across a
point where the runtime system might take over (for example, any
garbage collection point) or a non-tail call to another <a href="RSSA">RSSA</a>
function.  Those that are live go on the stack, while those that
aren&#8217;t live go into psuedo-registers.  From this information, we know
all we need to about each stack frame.  On the downside, nothing
further on is allowed to change this stack info; it is set in stone.</p></div>
</div>
</div>
</div>
<div id="footnotes"><hr></div>
<div id="footer">
<div id="footer-text">
</div>
<div id="footer-badges">
</div>
</div>
</body>
</html>
