494 lines
29 KiB
HTML
494 lines
29 KiB
HTML
<!DOCTYPE HTML>
|
||
<html lang="en" class="light sidebar-visible" dir="ltr">
|
||
<head>
|
||
<!-- Book generated using mdBook -->
|
||
<meta charset="UTF-8">
|
||
<title>The MIR (Mid-level IR) - Rust Compiler Development Guide</title>
|
||
|
||
|
||
<!-- Custom HTML head -->
|
||
|
||
<meta name="description" content="A guide to developing the Rust compiler (rustc)">
|
||
<meta name="viewport" content="width=device-width, initial-scale=1">
|
||
<meta name="theme-color" content="#ffffff">
|
||
|
||
<link rel="icon" href="../favicon.svg">
|
||
<link rel="shortcut icon" href="../favicon.png">
|
||
<link rel="stylesheet" href="../css/variables.css">
|
||
<link rel="stylesheet" href="../css/general.css">
|
||
<link rel="stylesheet" href="../css/chrome.css">
|
||
<link rel="stylesheet" href="../css/print.css" media="print">
|
||
|
||
<!-- Fonts -->
|
||
<link rel="stylesheet" href="../FontAwesome/css/font-awesome.css">
|
||
<link rel="stylesheet" href="../fonts/fonts.css">
|
||
|
||
<!-- Highlight.js Stylesheets -->
|
||
<link rel="stylesheet" id="highlight-css" href="../highlight.css">
|
||
<link rel="stylesheet" id="tomorrow-night-css" href="../tomorrow-night.css">
|
||
<link rel="stylesheet" id="ayu-highlight-css" href="../ayu-highlight.css">
|
||
|
||
<!-- Custom theme stylesheets -->
|
||
|
||
|
||
<!-- Provide site root and default themes to javascript -->
|
||
<script>
|
||
const path_to_root = "../";
|
||
const default_light_theme = "light";
|
||
const default_dark_theme = "navy";
|
||
</script>
|
||
<!-- Start loading toc.js asap -->
|
||
<script src="../toc.js"></script>
|
||
</head>
|
||
<body>
|
||
<div id="body-container">
|
||
<!-- Work around some values being stored in localStorage wrapped in quotes -->
|
||
<script>
|
||
try {
|
||
let theme = localStorage.getItem('mdbook-theme');
|
||
let sidebar = localStorage.getItem('mdbook-sidebar');
|
||
|
||
if (theme.startsWith('"') && theme.endsWith('"')) {
|
||
localStorage.setItem('mdbook-theme', theme.slice(1, theme.length - 1));
|
||
}
|
||
|
||
if (sidebar.startsWith('"') && sidebar.endsWith('"')) {
|
||
localStorage.setItem('mdbook-sidebar', sidebar.slice(1, sidebar.length - 1));
|
||
}
|
||
} catch (e) { }
|
||
</script>
|
||
|
||
<!-- Set the theme before any content is loaded, prevents flash -->
|
||
<script>
|
||
const default_theme = window.matchMedia("(prefers-color-scheme: dark)").matches ? default_dark_theme : default_light_theme;
|
||
let theme;
|
||
try { theme = localStorage.getItem('mdbook-theme'); } catch(e) { }
|
||
if (theme === null || theme === undefined) { theme = default_theme; }
|
||
const html = document.documentElement;
|
||
html.classList.remove('light')
|
||
html.classList.add(theme);
|
||
html.classList.add("js");
|
||
</script>
|
||
|
||
<input type="checkbox" id="sidebar-toggle-anchor" class="hidden">
|
||
|
||
<!-- Hide / unhide sidebar before it is displayed -->
|
||
<script>
|
||
let sidebar = null;
|
||
const sidebar_toggle = document.getElementById("sidebar-toggle-anchor");
|
||
if (document.body.clientWidth >= 1080) {
|
||
try { sidebar = localStorage.getItem('mdbook-sidebar'); } catch(e) { }
|
||
sidebar = sidebar || 'visible';
|
||
} else {
|
||
sidebar = 'hidden';
|
||
}
|
||
sidebar_toggle.checked = sidebar === 'visible';
|
||
html.classList.remove('sidebar-visible');
|
||
html.classList.add("sidebar-" + sidebar);
|
||
</script>
|
||
|
||
<nav id="sidebar" class="sidebar" aria-label="Table of contents">
|
||
<!-- populated by js -->
|
||
<mdbook-sidebar-scrollbox class="sidebar-scrollbox"></mdbook-sidebar-scrollbox>
|
||
<noscript>
|
||
<iframe class="sidebar-iframe-outer" src="../toc.html"></iframe>
|
||
</noscript>
|
||
<div id="sidebar-resize-handle" class="sidebar-resize-handle">
|
||
<div class="sidebar-resize-indicator"></div>
|
||
</div>
|
||
</nav>
|
||
|
||
<div id="page-wrapper" class="page-wrapper">
|
||
|
||
<div class="page">
|
||
<div id="menu-bar-hover-placeholder"></div>
|
||
<div id="menu-bar" class="menu-bar sticky">
|
||
<div class="left-buttons">
|
||
<label id="sidebar-toggle" class="icon-button" for="sidebar-toggle-anchor" title="Toggle Table of Contents" aria-label="Toggle Table of Contents" aria-controls="sidebar">
|
||
<i class="fa fa-bars"></i>
|
||
</label>
|
||
<button id="theme-toggle" class="icon-button" type="button" title="Change theme" aria-label="Change theme" aria-haspopup="true" aria-expanded="false" aria-controls="theme-list">
|
||
<i class="fa fa-paint-brush"></i>
|
||
</button>
|
||
<ul id="theme-list" class="theme-popup" aria-label="Themes" role="menu">
|
||
<li role="none"><button role="menuitem" class="theme" id="default_theme">Auto</button></li>
|
||
<li role="none"><button role="menuitem" class="theme" id="light">Light</button></li>
|
||
<li role="none"><button role="menuitem" class="theme" id="rust">Rust</button></li>
|
||
<li role="none"><button role="menuitem" class="theme" id="coal">Coal</button></li>
|
||
<li role="none"><button role="menuitem" class="theme" id="navy">Navy</button></li>
|
||
<li role="none"><button role="menuitem" class="theme" id="ayu">Ayu</button></li>
|
||
</ul>
|
||
<button id="search-toggle" class="icon-button" type="button" title="Search. (Shortkey: s)" aria-label="Toggle Searchbar" aria-expanded="false" aria-keyshortcuts="S" aria-controls="searchbar">
|
||
<i class="fa fa-search"></i>
|
||
</button>
|
||
</div>
|
||
|
||
<h1 class="menu-title">Rust Compiler Development Guide</h1>
|
||
|
||
<div class="right-buttons">
|
||
<a href="../print.html" title="Print this book" aria-label="Print this book">
|
||
<i id="print-button" class="fa fa-print"></i>
|
||
</a>
|
||
<a href="https://github.com/rust-lang/rustc-dev-guide" title="Git repository" aria-label="Git repository">
|
||
<i id="git-repository-button" class="fa fa-github"></i>
|
||
</a>
|
||
<a href="https://github.com/rust-lang/rustc-dev-guide/edit/master/src/mir/index.md" title="Suggest an edit" aria-label="Suggest an edit">
|
||
<i id="git-edit-button" class="fa fa-edit"></i>
|
||
</a>
|
||
|
||
</div>
|
||
</div>
|
||
|
||
<div id="search-wrapper" class="hidden">
|
||
<form id="searchbar-outer" class="searchbar-outer">
|
||
<input type="search" id="searchbar" name="searchbar" placeholder="Search this book ..." aria-controls="searchresults-outer" aria-describedby="searchresults-header">
|
||
</form>
|
||
<div id="searchresults-outer" class="searchresults-outer hidden">
|
||
<div id="searchresults-header" class="searchresults-header"></div>
|
||
<ul id="searchresults">
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
|
||
<!-- Apply ARIA attributes after the sidebar and the sidebar toggle button are added to the DOM -->
|
||
<script>
|
||
document.getElementById('sidebar-toggle').setAttribute('aria-expanded', sidebar === 'visible');
|
||
document.getElementById('sidebar').setAttribute('aria-hidden', sidebar !== 'visible');
|
||
Array.from(document.querySelectorAll('#sidebar a')).forEach(function(link) {
|
||
link.setAttribute('tabIndex', sidebar === 'visible' ? 0 : -1);
|
||
});
|
||
</script>
|
||
|
||
<div id="content" class="content">
|
||
<main>
|
||
<h1 id="the-mir-mid-level-ir"><a class="header" href="#the-mir-mid-level-ir">The MIR (Mid-level IR)</a></h1>
|
||
<ul>
|
||
<li><a href="#introduction-to-mir">Introduction to MIR</a></li>
|
||
<li><a href="#key-mir-vocabulary">Key MIR vocabulary</a></li>
|
||
<li><a href="#mir-data-types">MIR data types</a></li>
|
||
<li><a href="#representing-constants">Representing constants</a>
|
||
<ul>
|
||
<li><a href="#mir-constant-values">MIR constant values</a></li>
|
||
<li><a href="#valtrees">ValTrees</a></li>
|
||
<li><a href="#promoted-constants">Promoted constants</a></li>
|
||
</ul>
|
||
</li>
|
||
</ul>
|
||
<p>MIR is Rust's <em>Mid-level Intermediate Representation</em>. It is
|
||
constructed from <a href="../hir.html">HIR</a>. MIR was introduced in
|
||
<a href="https://rust-lang.github.io/rfcs/1211-mir.html">RFC 1211</a>. It is a radically simplified form of Rust that is used for
|
||
certain flow-sensitive safety checks – notably the borrow checker! –
|
||
and also for optimization and code generation.</p>
|
||
<p>If you'd like a very high-level introduction to MIR, as well as some
|
||
of the compiler concepts that it relies on (such as control-flow
|
||
graphs and desugaring), you may enjoy the
|
||
<a href="https://blog.rust-lang.org/2016/04/19/MIR.html">rust-lang blog post that introduced MIR</a>.</p>
|
||
<h2 id="introduction-to-mir"><a class="header" href="#introduction-to-mir">Introduction to MIR</a></h2>
|
||
<p>MIR is defined in the <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/index.html"><code>compiler/rustc_middle/src/mir/</code></a> module, but much of the code
|
||
that manipulates it is found in <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_mir_build/index.html"><code>compiler/rustc_mir_build</code></a>,
|
||
<a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_mir_transform/index.html"><code>compiler/rustc_mir_transform</code></a>, and
|
||
<a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_mir_dataflow/index.html"><code>compiler/rustc_mir_dataflow</code></a>.</p>
|
||
<p>Some of the key characteristics of MIR are:</p>
|
||
<ul>
|
||
<li>It is based on a <a href="../appendix/background.html#cfg">control-flow graph</a>.</li>
|
||
<li>It does not have nested expressions.</li>
|
||
<li>All types in MIR are fully explicit.</li>
|
||
</ul>
|
||
<h2 id="key-mir-vocabulary"><a class="header" href="#key-mir-vocabulary">Key MIR vocabulary</a></h2>
|
||
<p>This section introduces the key concepts of MIR, summarized here:</p>
|
||
<ul>
|
||
<li><strong>Basic blocks</strong>: units of the control-flow graph, consisting of:
|
||
<ul>
|
||
<li><strong>statements:</strong> actions with one successor</li>
|
||
<li><strong>terminators:</strong> actions with potentially multiple successors; always at
|
||
the end of a block</li>
|
||
<li>(if you're not familiar with the term <em>basic block</em>, see the <a href="../appendix/background.html#cfg">background
|
||
chapter</a>)</li>
|
||
</ul>
|
||
</li>
|
||
<li><strong>Locals:</strong> Memory locations allocated on the stack (conceptually, at
|
||
least), such as function arguments, local variables, and
|
||
temporaries. These are identified by an index, written with a
|
||
leading underscore, like <code>_1</code>. There is also a special "local"
|
||
(<code>_0</code>) allocated to store the return value.</li>
|
||
<li><strong>Places:</strong> expressions that identify a location in memory, like <code>_1</code> or
|
||
<code>_1.f</code>.</li>
|
||
<li><strong>Rvalues:</strong> expressions that produce a value. The "R" stands for
|
||
the fact that these are the "right-hand side" of an assignment.
|
||
<ul>
|
||
<li><strong>Operands:</strong> the arguments to an rvalue, which can either be a
|
||
constant (like <code>22</code>) or a place (like <code>_1</code>).</li>
|
||
</ul>
|
||
</li>
|
||
</ul>
|
||
<p>You can get a feeling for how MIR is constructed by translating simple
|
||
programs into MIR and reading the pretty printed output. In fact, the
|
||
playground makes this easy, since it supplies a MIR button that will
|
||
show you the MIR for your program. Try putting this program into play
|
||
(or <a href="https://play.rust-lang.org/?gist=30074856e62e74e91f06abd19bd72ece&version=stable&edition=2021">clicking on this link</a>), and then clicking the "MIR"
|
||
button on the top:</p>
|
||
<pre><pre class="playground"><code class="language-rust">fn main() {
|
||
let mut vec = Vec::new();
|
||
vec.push(1);
|
||
vec.push(2);
|
||
}</code></pre></pre>
|
||
<p>You should see something like:</p>
|
||
<pre><code class="language-mir">// WARNING: This output format is intended for human consumers only
|
||
// and is subject to change without notice. Knock yourself out.
|
||
fn main() -> () {
|
||
...
|
||
}
|
||
</code></pre>
|
||
<p>This is the MIR format for the <code>main</code> function.
|
||
MIR shown by above link is optimized.
|
||
Some statements like <code>StorageLive</code> are removed in optimization.
|
||
This happens because the compiler notices the value is never accessed in the code.
|
||
We can use <code>rustc [filename].rs -Z mir-opt-level=0 --emit mir</code> to view unoptimized MIR.
|
||
This requires the nightly toolchain.</p>
|
||
<p><strong>Variable declarations.</strong> If we drill in a bit, we'll see it begins
|
||
with a bunch of variable declarations. They look like this:</p>
|
||
<pre><code class="language-mir">let mut _0: (); // return place
|
||
let mut _1: std::vec::Vec<i32>; // in scope 0 at src/main.rs:2:9: 2:16
|
||
let mut _2: ();
|
||
let mut _3: &mut std::vec::Vec<i32>;
|
||
let mut _4: ();
|
||
let mut _5: &mut std::vec::Vec<i32>;
|
||
</code></pre>
|
||
<p>You can see that variables in MIR don't have names, they have indices,
|
||
like <code>_0</code> or <code>_1</code>. We also intermingle the user's variables (e.g.,
|
||
<code>_1</code>) with temporary values (e.g., <code>_2</code> or <code>_3</code>). You can tell apart
|
||
user-defined variables because they have debuginfo associated to them (see below).</p>
|
||
<p><strong>User variable debuginfo.</strong> Below the variable declarations, we find the only
|
||
hint that <code>_1</code> represents a user variable:</p>
|
||
<pre><code class="language-mir">scope 1 {
|
||
debug vec => _1; // in scope 1 at src/main.rs:2:9: 2:16
|
||
}
|
||
</code></pre>
|
||
<p>Each <code>debug <Name> => <Place>;</code> annotation describes a named user variable,
|
||
and where (i.e. the place) a debugger can find the data of that variable.
|
||
Here the mapping is trivial, but optimizations may complicate the place,
|
||
or lead to multiple user variables sharing the same place.
|
||
Additionally, closure captures are described using the same system, and so
|
||
they're complicated even without optimizations, e.g.: <code>debug x => (*((*_1).0: &T));</code>.</p>
|
||
<p>The "scope" blocks (e.g., <code>scope 1 { .. }</code>) describe the lexical structure of
|
||
the source program (which names were in scope when), so any part of the program
|
||
annotated with <code>// in scope 0</code> would be missing <code>vec</code>, if you were stepping
|
||
through the code in a debugger, for example.</p>
|
||
<p><strong>Basic blocks.</strong> Reading further, we see our first <strong>basic block</strong> (naturally
|
||
it may look slightly different when you view it, and I am ignoring some of the
|
||
comments):</p>
|
||
<pre><code class="language-mir">bb0: {
|
||
StorageLive(_1);
|
||
_1 = const <std::vec::Vec<T>>::new() -> bb2;
|
||
}
|
||
</code></pre>
|
||
<p>A basic block is defined by a series of <strong>statements</strong> and a final
|
||
<strong>terminator</strong>. In this case, there is one statement:</p>
|
||
<pre><code class="language-mir">StorageLive(_1);
|
||
</code></pre>
|
||
<p>This statement indicates that the variable <code>_1</code> is "live", meaning
|
||
that it may be used later – this will persist until we encounter a
|
||
<code>StorageDead(_1)</code> statement, which indicates that the variable <code>_1</code> is
|
||
done being used. These "storage statements" are used by LLVM to
|
||
allocate stack space.</p>
|
||
<p>The <strong>terminator</strong> of the block <code>bb0</code> is the call to <code>Vec::new</code>:</p>
|
||
<pre><code class="language-mir">_1 = const <std::vec::Vec<T>>::new() -> bb2;
|
||
</code></pre>
|
||
<p>Terminators are different from statements because they can have more
|
||
than one successor – that is, control may flow to different
|
||
places. Function calls like the call to <code>Vec::new</code> are always
|
||
terminators because of the possibility of unwinding, although in the
|
||
case of <code>Vec::new</code> we are able to see that indeed unwinding is not
|
||
possible, and hence we list only one successor block, <code>bb2</code>.</p>
|
||
<p>If we look ahead to <code>bb2</code>, we will see it looks like this:</p>
|
||
<pre><code class="language-mir">bb2: {
|
||
StorageLive(_3);
|
||
_3 = &mut _1;
|
||
_2 = const <std::vec::Vec<T>>::push(move _3, const 1i32) -> [return: bb3, unwind: bb4];
|
||
}
|
||
</code></pre>
|
||
<p>Here there are two statements: another <code>StorageLive</code>, introducing the <code>_3</code>
|
||
temporary, and then an assignment:</p>
|
||
<pre><code class="language-mir">_3 = &mut _1;
|
||
</code></pre>
|
||
<p>Assignments in general have the form:</p>
|
||
<pre><code class="language-text"><Place> = <Rvalue>
|
||
</code></pre>
|
||
<p>A place is an expression like <code>_3</code>, <code>_3.f</code> or <code>*_3</code> – it denotes a
|
||
location in memory. An <strong>Rvalue</strong> is an expression that creates a
|
||
value: in this case, the rvalue is a mutable borrow expression, which
|
||
looks like <code>&mut <Place></code>. So we can kind of define a grammar for
|
||
rvalues like so:</p>
|
||
<pre><code class="language-text"><Rvalue> = & (mut)? <Place>
|
||
| <Operand> + <Operand>
|
||
| <Operand> - <Operand>
|
||
| ...
|
||
|
||
<Operand> = Constant
|
||
| copy Place
|
||
| move Place
|
||
</code></pre>
|
||
<p>As you can see from this grammar, rvalues cannot be nested – they can
|
||
only reference places and constants. Moreover, when you use a place,
|
||
we indicate whether we are <strong>copying it</strong> (which requires that the
|
||
place have a type <code>T</code> where <code>T: Copy</code>) or <strong>moving it</strong> (which works
|
||
for a place of any type). So, for example, if we had the expression <code>x = a + b + c</code> in Rust, that would get compiled to two statements and a
|
||
temporary:</p>
|
||
<pre><code class="language-mir">TMP1 = a + b
|
||
x = TMP1 + c
|
||
</code></pre>
|
||
<p>(<a href="https://play.rust-lang.org/?gist=1751196d63b2a71f8208119e59d8a5b6&version=stable">Try it and see</a>, though you may want to do release mode to skip
|
||
over the overflow checks.)</p>
|
||
<h2 id="mir-data-types"><a class="header" href="#mir-data-types">MIR data types</a></h2>
|
||
<p>The MIR data types are defined in the <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/index.html"><code>compiler/rustc_middle/src/mir/</code></a>
|
||
module. Each of the key concepts mentioned in the previous section
|
||
maps in a fairly straightforward way to a Rust type.</p>
|
||
<p>The main MIR data type is <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/struct.Body.html"><code>Body</code></a>. It contains the data for a single
|
||
function (along with sub-instances of Mir for "promoted constants",
|
||
but <a href="#promoted">you can read about those below</a>).</p>
|
||
<ul>
|
||
<li><strong>Basic blocks</strong>: The basic blocks are stored in the field
|
||
<a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/struct.Body.html#structfield.basic_blocks"><code>Body::basic_blocks</code></a>; this is a vector
|
||
of <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/struct.BasicBlockData.html"><code>BasicBlockData</code></a> structures. Nobody ever references a
|
||
basic block directly: instead, we pass around <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/struct.BasicBlock.html"><code>BasicBlock</code></a>
|
||
values, which are <a href="../appendix/glossary.html#newtype">newtype'd</a> indices into this vector.</li>
|
||
<li><strong>Statements</strong> are represented by the type <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/struct.Statement.html"><code>Statement</code></a>.</li>
|
||
<li><strong>Terminators</strong> are represented by the <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/terminator/struct.Terminator.html"><code>Terminator</code></a>.</li>
|
||
<li><strong>Locals</strong> are represented by a <a href="../appendix/glossary.html#newtype">newtype'd</a> index type <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/struct.Local.html"><code>Local</code></a>.
|
||
The data for a local variable is found in the
|
||
<a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/struct.Body.html#structfield.local_decls"><code>Body::local_decls</code></a> vector. There is also a special constant
|
||
<a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/constant.RETURN_PLACE.html"><code>RETURN_PLACE</code></a> identifying the special "local" representing the return value.</li>
|
||
<li><strong>Places</strong> are identified by the struct <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/struct.Place.html"><code>Place</code></a>. There are a few
|
||
fields:
|
||
<ul>
|
||
<li>Local variables like <code>_1</code></li>
|
||
<li><strong>Projections</strong>, which are fields or other things that "project
|
||
out" from a base place. These are represented by the <a href="../appendix/glossary.html#newtype">newtype'd</a> type
|
||
<a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/enum.ProjectionElem.html"><code>ProjectionElem</code></a>. So e.g. the place <code>_1.f</code> is a projection,
|
||
with <code>f</code> being the "projection element" and <code>_1</code> being the base
|
||
path. <code>*_1</code> is also a projection, with the <code>*</code> being represented
|
||
by the <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/enum.ProjectionElem.html#variant.Deref"><code>ProjectionElem::Deref</code></a> element.</li>
|
||
</ul>
|
||
</li>
|
||
<li><strong>Rvalues</strong> are represented by the enum <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/enum.Rvalue.html"><code>Rvalue</code></a>.</li>
|
||
<li><strong>Operands</strong> are represented by the enum <a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/enum.Operand.html"><code>Operand</code></a>.</li>
|
||
</ul>
|
||
<h2 id="representing-constants"><a class="header" href="#representing-constants">Representing constants</a></h2>
|
||
<p>When code has reached the MIR stage, constants can generally come in two forms:
|
||
<em>MIR constants</em> (<a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/enum.Const.html"><code>mir::Constant</code></a>) and <em>type system constants</em> (<a href="https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/ty/struct.Const.html"><code>ty::Const</code></a>).
|
||
MIR constants are used as operands: in <code>x + CONST</code>, <code>CONST</code> is a MIR constant;
|
||
similarly, in <code>x + 2</code>, <code>2</code> is a MIR constant. Type system constants are used in
|
||
the type system, in particular for array lengths but also for const generics.</p>
|
||
<p>Generally, both kinds of constants can be "unevaluated" or "already evaluated".
|
||
An unevaluated constant simply stores the <code>DefId</code> of what needs to be evaluated
|
||
to compute this result. An evaluated constant (a "value") has already been
|
||
computed; their representation differs between type system constants and MIR
|
||
constants: MIR constants evaluate to a <code>mir::ConstValue</code>; type system constants
|
||
evaluate to a <code>ty::ValTree</code>.</p>
|
||
<p>Type system constants have some more variants to support const generics: they
|
||
can refer to local const generic parameters, and they are subject to inference.
|
||
Furthermore, the <code>mir::Constant::Ty</code> variant lets us use an arbitrary type
|
||
system constant as a MIR constant; this happens whenever a const generic
|
||
parameter is used as an operand.</p>
|
||
<h3 id="mir-constant-values"><a class="header" href="#mir-constant-values">MIR constant values</a></h3>
|
||
<p>In general, a MIR constant value (<code>mir::ConstValue</code>) was computed by evaluating
|
||
some constant the user wrote. This <a href="../const-eval.html">const evaluation</a> produces
|
||
a very low-level representation of the result in terms of individual bytes. We
|
||
call this an "indirect" constant (<code>mir::ConstValue::Indirect</code>) since the value
|
||
is stored in-memory.</p>
|
||
<p>However, storing everything in-memory would be awfully inefficient. Hence there
|
||
are some other variants in <code>mir::ConstValue</code> that can represent certain simple
|
||
and common values more efficiently. In particular, everything that can be
|
||
directly written as a literal in Rust (integers, floats, chars, bools, but also
|
||
<code>"string literals"</code> and <code>b"byte string literals"</code>) has an optimized variant that
|
||
avoids the full overhead of the in-memory representation.</p>
|
||
<h3 id="valtrees"><a class="header" href="#valtrees">ValTrees</a></h3>
|
||
<p>An evaluated type system constant is a "valtree". The <code>ty::ValTree</code> datastructure
|
||
allows us to represent</p>
|
||
<ul>
|
||
<li>arrays,</li>
|
||
<li>many structs,</li>
|
||
<li>tuples,</li>
|
||
<li>enums and,</li>
|
||
<li>most primitives.</li>
|
||
</ul>
|
||
<p>The most important rule for
|
||
this representation is that every value must be uniquely represented. In other
|
||
words: a specific value must only be representable in one specific way. For example: there is only
|
||
one way to represent an array of two integers as a <code>ValTree</code>:
|
||
<code>Branch([Leaf(first_int), Leaf(second_int)])</code>.
|
||
Even though theoretically a <code>[u32; 2]</code> could be encoded in a <code>u64</code> and thus just be a
|
||
<code>Leaf(bits_of_two_u32)</code>, that is not a legal construction of <code>ValTree</code>
|
||
(and is very complex to do, so it is unlikely anyone is tempted to do so).</p>
|
||
<p>These rules also mean that some values are not representable. There can be no <code>union</code>s in type
|
||
level constants, as it is not clear how they should be represented, because their active variant
|
||
is unknown. Similarly there is no way to represent raw pointers, as addresses are unknown at
|
||
compile-time and thus we cannot make any assumptions about them. References on the other hand
|
||
<em>can</em> be represented, as equality for references is defined as equality on their value, so we
|
||
ignore their address and just look at the backing value. We must make sure that the pointer values
|
||
of the references are not observable at compile time. We thus encode <code>&42</code> exactly like <code>42</code>.
|
||
Any conversion from
|
||
valtree back to a MIR constant value must reintroduce an actual indirection. At codegen time the
|
||
addresses may be deduplicated between multiple uses or not, entirely depending on arbitrary
|
||
optimization choices.</p>
|
||
<p>As a consequence, all decoding of <code>ValTree</code> must happen by matching on the type first and making
|
||
decisions depending on that. The value itself gives no useful information without the type that
|
||
belongs to it.</p>
|
||
<p><a id="promoted"></a></p>
|
||
<h3 id="promoted-constants"><a class="header" href="#promoted-constants">Promoted constants</a></h3>
|
||
<p>See the const-eval WG's <a href="https://github.com/rust-lang/const-eval/blob/master/promotion.md">docs on promotion</a>.</p>
|
||
|
||
</main>
|
||
|
||
<nav class="nav-wrapper" aria-label="Page navigation">
|
||
<!-- Mobile navigation buttons -->
|
||
<a rel="prev" href="../thir.html" class="mobile-nav-chapters previous" title="Previous chapter" aria-label="Previous chapter" aria-keyshortcuts="Left">
|
||
<i class="fa fa-angle-left"></i>
|
||
</a>
|
||
|
||
<a rel="next prefetch" href="../mir/construction.html" class="mobile-nav-chapters next" title="Next chapter" aria-label="Next chapter" aria-keyshortcuts="Right">
|
||
<i class="fa fa-angle-right"></i>
|
||
</a>
|
||
|
||
<div style="clear: both"></div>
|
||
</nav>
|
||
</div>
|
||
</div>
|
||
|
||
<nav class="nav-wide-wrapper" aria-label="Page navigation">
|
||
<a rel="prev" href="../thir.html" class="nav-chapters previous" title="Previous chapter" aria-label="Previous chapter" aria-keyshortcuts="Left">
|
||
<i class="fa fa-angle-left"></i>
|
||
</a>
|
||
|
||
<a rel="next prefetch" href="../mir/construction.html" class="nav-chapters next" title="Next chapter" aria-label="Next chapter" aria-keyshortcuts="Right">
|
||
<i class="fa fa-angle-right"></i>
|
||
</a>
|
||
</nav>
|
||
|
||
</div>
|
||
|
||
|
||
|
||
|
||
<script>
|
||
window.playground_copyable = true;
|
||
</script>
|
||
|
||
|
||
<script src="../elasticlunr.min.js"></script>
|
||
<script src="../mark.min.js"></script>
|
||
<script src="../searcher.js"></script>
|
||
|
||
<script src="../clipboard.min.js"></script>
|
||
<script src="../highlight.js"></script>
|
||
<script src="../book.js"></script>
|
||
|
||
<!-- Custom JS scripts -->
|
||
<script src="../mermaid.min.js"></script>
|
||
<script src="../mermaid-init.js"></script>
|
||
|
||
|
||
</div>
|
||
</body>
|
||
</html>
|