Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / d46ba26a6b513086ad575e5db0069b03d41fca13 / . / docs / GarbageCollection.html
blob: 761e1d08cafc942dd713f7b67baa1ca4ddf9b8db [file] [log] [blame]
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]5 <meta http-equiv="Content-Type" Content="text/html; charset=UTF-8" >
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]6 <title>Accurate Garbage Collection with LLVM</title>
7 <link rel="stylesheet" href="llvm.css" type="text/css">
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]8 <style type="text/css">
9 .rowhead { text-align: left; background: inherit; }
10 .indent { padding-left: 1em; }
11 .optl { color: #BFBFBF; }
12 </style>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]13</head>
14<body>
15
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]16<h1>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]17 Accurate Garbage Collection with LLVM
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]18</h1>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]19
20<ol>
21 <li><a href="#introduction">Introduction</a>
22 <ul>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]23 <li><a href="#feature">Goals and non-goals</a></li>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]24 </ul>
25 </li>
26
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]27 <li><a href="#quickstart">Getting started</a>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]28 <ul>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]29 <li><a href="#quickstart-compiler">In your compiler</a></li>
30 <li><a href="#quickstart-runtime">In your runtime library</a></li>
31 <li><a href="#shadow-stack">About the shadow stack</a></li>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]32 </ul>
33 </li>
34
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]35 <li><a href="#core">Core support</a>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]36 <ul>
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]37 <li><a href="#gcattr">Specifying GC code generation:
38 <tt>gc "..."</tt></a></li>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]39 <li><a href="#gcroot">Identifying GC roots on the stack:
40 <tt>llvm.gcroot</tt></a></li>
41 <li><a href="#barriers">Reading and writing references in the heap</a>
42 <ul>
43 <li><a href="#gcwrite">Write barrier: <tt>llvm.gcwrite</tt></a></li>
44 <li><a href="#gcread">Read barrier: <tt>llvm.gcread</tt></a></li>
45 </ul>
46 </li>
47 </ul>
48 </li>
49
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]50 <li><a href="#plugin">Compiler plugin interface</a>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]51 <ul>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]52 <li><a href="#collector-algos">Overview of available features</a></li>
53 <li><a href="#stack-map">Computing stack maps</a></li>
54 <li><a href="#init-roots">Initializing roots to null:
55 <tt>InitRoots</tt></a></li>
56 <li><a href="#custom">Custom lowering of intrinsics: <tt>CustomRoots</tt>,
57 <tt>CustomReadBarriers</tt>, and <tt>CustomWriteBarriers</tt></a></li>
58 <li><a href="#safe-points">Generating safe points:
59 <tt>NeededSafePoints</tt></a></li>
60 <li><a href="#assembly">Emitting assembly code:
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]61 <tt>GCMetadataPrinter</tt></a></li>
Chris Lattner0b02dbc2004-07-09 05:03:54 +0000[diff] [blame]62 </ul>
Chris Lattner9b2a1842004-05-27 05:52:10 +0000[diff] [blame]63 </li>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]64
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]65 <li><a href="#runtime-impl">Implementing a collector runtime</a>
66 <ul>
67 <li><a href="#gcdescriptors">Tracing GC pointers from heap
68 objects</a></li>
69 </ul>
70 </li>
71
72 <li><a href="#references">References</a></li>
73
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]74</ol>
75
76<div class="doc_author">
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]77 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a> and
78 Gordon Henriksen</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]79</div>
80
81<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]82<h2>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]83 <a name="introduction">Introduction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]84</h2>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]85<!-- *********************************************************************** -->
86
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]87<div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]88
89<p>Garbage collection is a widely used technique that frees the programmer from
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]90having to know the lifetimes of heap objects, making software easier to produce
91and maintain. Many programming languages rely on garbage collection for
92automatic memory management. There are two primary forms of garbage collection:
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]93conservative and accurate.</p>
94
95<p>Conservative garbage collection often does not require any special support
96from either the language or the compiler: it can handle non-type-safe
97programming languages (such as C/C++) and does not require any special
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]98information from the compiler. The
Jeff Cohen65fc36b2007-04-18 17:26:14 +0000[diff] [blame]99<a href="http://www.hpl.hp.com/personal/Hans_Boehm/gc/">Boehm collector</a> is
100an example of a state-of-the-art conservative collector.</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]101
102<p>Accurate garbage collection requires the ability to identify all pointers in
103the program at run-time (which requires that the source-language be type-safe in
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]104most cases). Identifying pointers at run-time requires compiler support to
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]105locate all places that hold live pointer variables at run-time, including the
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]106<a href="#gcroot">processor stack and registers</a>.</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]107
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]108<p>Conservative garbage collection is attractive because it does not require any
109special compiler support, but it does have problems. In particular, because the
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]110conservative garbage collector cannot <i>know</i> that a particular word in the
111machine is a pointer, it cannot move live objects in the heap (preventing the
112use of compacting and generational GC algorithms) and it can occasionally suffer
113from memory leaks due to integer values that happen to point to objects in the
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]114program. In addition, some aggressive compiler transformations can break
115conservative garbage collectors (though these seem rare in practice).</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]116
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]117<p>Accurate garbage collectors do not suffer from any of these problems, but
118they can suffer from degraded scalar optimization of the program. In particular,
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]119because the runtime must be able to identify and update all pointers active in
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]120the program, some optimizations are less effective. In practice, however, the
Chris Lattner7f3b36d2009-05-13 18:02:09 +0000[diff] [blame]121locality and performance benefits of using aggressive garbage collection
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]122techniques dominates any low-level losses.</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]123
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]124<p>This document describes the mechanisms and interfaces provided by LLVM to
125support accurate garbage collection.</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]126
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]127<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]128<h3>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]129 <a name="feature">Goals and non-goals</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]130</h3>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]131
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]132<div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]133
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]134<p>LLVM's intermediate representation provides <a href="#intrinsics">garbage
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]135collection intrinsics</a> that offer support for a broad class of
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]136collector models. For instance, the intrinsics permit:</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]137
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]138<ul>
139 <li>semi-space collectors</li>
140 <li>mark-sweep collectors</li>
141 <li>generational collectors</li>
142 <li>reference counting</li>
143 <li>incremental collectors</li>
144 <li>concurrent collectors</li>
145 <li>cooperative collectors</li>
146</ul>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]147
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]148<p>We hope that the primitive support built into the LLVM IR is sufficient to
149support a broad class of garbage collected languages including Scheme, ML, Java,
150C#, Perl, Python, Lua, Ruby, other scripting languages, and more.</p>
151
NAKAMURA Takumi4d6deb02011-04-09 09:51:57 +0000[diff] [blame]152<p>However, LLVM does not itself provide a garbage collector&mdash;this should
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]153be part of your language's runtime library. LLVM provides a framework for
154compile time <a href="#plugin">code generation plugins</a>. The role of these
155plugins is to generate code and data structures which conforms to the <em>binary
156interface</em> specified by the <em>runtime library</em>. This is similar to the
157relationship between LLVM and DWARF debugging info, for example. The
158difference primarily lies in the lack of an established standard in the domain
NAKAMURA Takumi4d6deb02011-04-09 09:51:57 +0000[diff] [blame]159of garbage collection&mdash;thus the plugins.</p>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]160
161<p>The aspects of the binary interface with which LLVM's GC support is
162concerned are:</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]163
164<ul>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]165 <li>Creation of GC-safe points within code where collection is allowed to
166 execute safely.</li>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]167 <li>Computation of the stack map. For each safe point in the code, object
168 references within the stack frame must be identified so that the
169 collector may traverse and perhaps update them.</li>
170 <li>Write barriers when storing object references to the heap. These are
171 commonly used to optimize incremental scans in generational
172 collectors.</li>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]173 <li>Emission of read barriers when loading object references. These are
174 useful for interoperating with concurrent collectors.</li>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]175</ul>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]176
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]177<p>There are additional areas that LLVM does not directly address:</p>
178
179<ul>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]180 <li>Registration of global roots with the runtime.</li>
181 <li>Registration of stack map entries with the runtime.</li>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]182 <li>The functions used by the program to allocate memory, trigger a
183 collection, etc.</li>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]184 <li>Computation or compilation of type maps, or registration of them with
185 the runtime. These are used to crawl the heap for object
186 references.</li>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]187</ul>
188
189<p>In general, LLVM's support for GC does not include features which can be
190adequately addressed with other features of the IR and does not specify a
191particular binary interface. On the plus side, this means that you should be
192able to integrate LLVM with an existing runtime. On the other hand, it leaves
193a lot of work for the developer of a novel language. However, it's easy to get
194started quickly and scale up to a more sophisticated implementation as your
195compiler matures.</p>
196
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]197</div>
198
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]199</div>
200
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]201<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]202<h2>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]203 <a name="quickstart">Getting started</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]204</h2>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]205<!-- *********************************************************************** -->
206
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]207<div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]208
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]209<p>Using a GC with LLVM implies many things, for example:</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]210
211<ul>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]212 <li>Write a runtime library or find an existing one which implements a GC
213 heap.<ol>
214 <li>Implement a memory allocator.</li>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]215 <li>Design a binary interface for the stack map, used to identify
216 references within a stack frame on the machine stack.*</li>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]217 <li>Implement a stack crawler to discover functions on the call stack.*</li>
218 <li>Implement a registry for global roots.</li>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]219 <li>Design a binary interface for type maps, used to identify references
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]220 within heap objects.</li>
221 <li>Implement a collection routine bringing together all of the above.</li>
222 </ol></li>
223 <li>Emit compatible code from your compiler.<ul>
224 <li>Initialization in the main function.</li>
225 <li>Use the <tt>gc "..."</tt> attribute to enable GC code generation
226 (or <tt>F.setGC("...")</tt>).</li>
227 <li>Use <tt>@llvm.gcroot</tt> to mark stack roots.</li>
228 <li>Use <tt>@llvm.gcread</tt> and/or <tt>@llvm.gcwrite</tt> to
229 manipulate GC references, if necessary.</li>
230 <li>Allocate memory using the GC allocation routine provided by the
231 runtime library.</li>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]232 <li>Generate type maps according to your runtime's binary interface.</li>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]233 </ul></li>
234 <li>Write a compiler plugin to interface LLVM with the runtime library.*<ul>
235 <li>Lower <tt>@llvm.gcread</tt> and <tt>@llvm.gcwrite</tt> to appropriate
236 code sequences.*</li>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]237 <li>Compile LLVM's stack map to the binary form expected by the
238 runtime.</li>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]239 </ul></li>
240 <li>Load the plugin into the compiler. Use <tt>llc -load</tt> or link the
241 plugin statically with your language's compiler.*</li>
242 <li>Link program executables with the runtime.</li>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]243</ul>
244
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]245<p>To help with several of these tasks (those indicated with a *), LLVM
246includes a highly portable, built-in ShadowStack code generator. It is compiled
247into <tt>llc</tt> and works even with the interpreter and C backends.</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]248
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]249<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]250<h3>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]251 <a name="quickstart-compiler">In your compiler</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]252</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]253
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]254<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]255
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]256<p>To turn the shadow stack on for your functions, first call:</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]257
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]258<div class="doc_code"><pre
259>F.setGC("shadow-stack");</pre></div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]260
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]261<p>for each function your compiler emits. Since the shadow stack is built into
262LLVM, you do not need to load a plugin.</p>
263
264<p>Your compiler must also use <tt>@llvm.gcroot</tt> as documented.
265Don't forget to create a root for each intermediate value that is generated
266when evaluating an expression. In <tt>h(f(), g())</tt>, the result of
267<tt>f()</tt> could easily be collected if evaluating <tt>g()</tt> triggers a
268collection.</p>
269
270<p>There's no need to use <tt>@llvm.gcread</tt> and <tt>@llvm.gcwrite</tt> over
271plain <tt>load</tt> and <tt>store</tt> for now. You will need them when
272switching to a more advanced GC.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]273
274</div>
275
276<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]277<h3>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]278 <a name="quickstart-runtime">In your runtime</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]279</h3>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]280
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]281<div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]282
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]283<p>The shadow stack doesn't imply a memory allocation algorithm. A semispace
284collector or building atop <tt>malloc</tt> are great places to start, and can
285be implemented with very little code.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]286
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]287<p>When it comes time to collect, however, your runtime needs to traverse the
288stack roots, and for this it needs to integrate with the shadow stack. Luckily,
289doing so is very simple. (This code is heavily commented to help you
290understand the data structure, but there are only 20 lines of meaningful
291code.)</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]292
293</div>
294
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]295<div class="doc_code"><pre
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]296>/// @brief The map for a single function's stack frame. One of these is
297/// compiled as constant data into the executable for each function.
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]298///
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]299/// Storage of metadata values is elided if the %metadata parameter to
300/// @llvm.gcroot is null.
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]301struct FrameMap {
302 int32_t NumRoots; //&lt; Number of roots in stack frame.
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]303 int32_t NumMeta; //&lt; Number of metadata entries. May be &lt; NumRoots.
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]304 const void *Meta[0]; //&lt; Metadata for each root.
305};
306
307/// @brief A link in the dynamic shadow stack. One of these is embedded in the
308/// stack frame of each function on the call stack.
309struct StackEntry {
310 StackEntry *Next; //&lt; Link to next stack entry (the caller's).
311 const FrameMap *Map; //&lt; Pointer to constant FrameMap.
312 void *Roots[0]; //&lt; Stack roots (in-place array).
313};
314
315/// @brief The head of the singly-linked list of StackEntries. Functions push
316/// and pop onto this in their prologue and epilogue.
317///
318/// Since there is only a global list, this technique is not threadsafe.
319StackEntry *llvm_gc_root_chain;
320
321/// @brief Calls Visitor(root, meta) for each GC root on the stack.
322/// root and meta are exactly the values passed to
323/// <tt>@llvm.gcroot</tt>.
324///
325/// Visitor could be a function to recursively mark live objects. Or it
326/// might copy them to another heap or generation.
327///
328/// @param Visitor A function to invoke for every GC root on the stack.
329void visitGCRoots(void (*Visitor)(void **Root, const void *Meta)) {
330 for (StackEntry *R = llvm_gc_root_chain; R; R = R->Next) {
331 unsigned i = 0;
332
333 // For roots [0, NumMeta), the metadata pointer is in the FrameMap.
334 for (unsigned e = R->Map->NumMeta; i != e; ++i)
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]335 Visitor(&amp;R->Roots[i], R->Map->Meta[i]);
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]336
337 // For roots [NumMeta, NumRoots), the metadata pointer is null.
338 for (unsigned e = R->Map->NumRoots; i != e; ++i)
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]339 Visitor(&amp;R->Roots[i], NULL);
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]340 }
341}</pre></div>
342
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]343<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]344<h3>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]345 <a name="shadow-stack">About the shadow stack</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]346</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]347
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]348<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]349
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]350<p>Unlike many GC algorithms which rely on a cooperative code generator to
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]351compile stack maps, this algorithm carefully maintains a linked list of stack
352roots [<a href="#henderson02">Henderson2002</a>]. This so-called "shadow stack"
353mirrors the machine stack. Maintaining this data structure is slower than using
354a stack map compiled into the executable as constant data, but has a significant
355portability advantage because it requires no special support from the target
356code generator, and does not require tricky platform-specific code to crawl
357the machine stack.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]358
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]359<p>The tradeoff for this simplicity and portability is:</p>
360
361<ul>
362 <li>High overhead per function call.</li>
363 <li>Not thread-safe.</li>
364</ul>
365
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]366<p>Still, it's an easy way to get started. After your compiler and runtime are
367up and running, writing a <a href="#plugin">plugin</a> will allow you to take
368advantage of <a href="#collector-algos">more advanced GC features</a> of LLVM
369in order to improve performance.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]370
371</div>
372
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]373</div>
374
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]375<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]376<h2>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]377 <a name="core">IR features</a><a name="intrinsics"></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]378</h2>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]379<!-- *********************************************************************** -->
380
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]381<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]382
383<p>This section describes the garbage collection facilities provided by the
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]384<a href="LangRef.html">LLVM intermediate representation</a>. The exact behavior
385of these IR features is specified by the binary interface implemented by a
386<a href="#plugin">code generation plugin</a>, not by this document.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]387
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]388<p>These facilities are limited to those strictly necessary; they are not
389intended to be a complete interface to any garbage collector. A program will
390need to interface with the GC library using the facilities provided by that
391program.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]392
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]393<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]394<h3>
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]395 <a name="gcattr">Specifying GC code generation: <tt>gc "..."</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]396</h3>
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]397
398<div class="doc_code"><tt>
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]399 define <i>ty</i> @<i>name</i>(...) <span style="text-decoration: underline">gc "<i>name</i>"</span> { ...
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]400</tt></div>
401
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]402<div>
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]403
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]404<p>The <tt>gc</tt> function attribute is used to specify the desired GC style
405to the compiler. Its programmatic equivalent is the <tt>setGC</tt> method of
406<tt>Function</tt>.</p>
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]407
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]408<p>Setting <tt>gc "<i>name</i>"</tt> on a function triggers a search for a
409matching code generation plugin "<i>name</i>"; it is that plugin which defines
410the exact nature of the code generated to support GC. If none is found, the
411compiler will raise an error.</p>
412
413<p>Specifying the GC style on a per-function basis allows LLVM to link together
414programs that use different garbage collection algorithms (or none at all).</p>
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]415
416</div>
417
418<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]419<h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]420 <a name="gcroot">Identifying GC roots on the stack: <tt>llvm.gcroot</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]421</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]422
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]423<div class="doc_code"><tt>
Chris Lattner17340552008-04-24 06:00:30 +0000[diff] [blame]424 void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]425</tt></div>
426
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]427<div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]428
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]429<p>The <tt>llvm.gcroot</tt> intrinsic is used to inform LLVM that a stack
430variable references an object on the heap and is to be tracked for garbage
431collection. The exact impact on generated code is specified by a <a
432href="#plugin">compiler plugin</a>.</p>
433
434<p>A compiler which uses mem2reg to raise imperative code using <tt>alloca</tt>
435into SSA form need only add a call to <tt>@llvm.gcroot</tt> for those variables
436which a pointers into the GC heap.</p>
437
438<p>It is also important to mark intermediate values with <tt>llvm.gcroot</tt>.
439For example, consider <tt>h(f(), g())</tt>. Beware leaking the result of
440<tt>f()</tt> in the case that <tt>g()</tt> triggers a collection.</p>
441
442<p>The first argument <b>must</b> be a value referring to an alloca instruction
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]443or a bitcast of an alloca. The second contains a pointer to metadata that
444should be associated with the pointer, and <b>must</b> be a constant or global
445value address. If your target collector uses tags, use a null pointer for
446metadata.</p>
447
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]448<p>The <tt>%metadata</tt> argument can be used to avoid requiring heap objects
449to have 'isa' pointers or tag bits. [<a href="#appel89">Appel89</a>, <a
450href="#goldberg91">Goldberg91</a>, <a href="#tolmach94">Tolmach94</a>] If
451specified, its value will be tracked along with the location of the pointer in
452the stack frame.</p>
453
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]454<p>Consider the following fragment of Java code:</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]455
456<pre>
457 {
458 Object X; // A null-initialized reference to an object
459 ...
460 }
461</pre>
462
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]463<p>This block (which may be located in the middle of a function or in a loop
464nest), could be compiled to this LLVM code:</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]465
466<pre>
467Entry:
468 ;; In the entry block for the function, allocate the
469 ;; stack space for X, which is an LLVM pointer.
470 %X = alloca %Object*
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]471
472 ;; Tell LLVM that the stack space is a stack root.
473 ;; Java has type-tags on objects, so we pass null as metadata.
474 %tmp = bitcast %Object** %X to i8**
Chris Lattner17340552008-04-24 06:00:30 +0000[diff] [blame]475 call void @llvm.gcroot(i8** %X, i8* null)
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]476 ...
477
478 ;; "CodeBlock" is the block corresponding to the start
Reid Spencer03d186a2004-05-25 08:45:31 +0000[diff] [blame]479 ;; of the scope above.
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]480CodeBlock:
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]481 ;; Java null-initializes pointers.
482 store %Object* null, %Object** %X
483
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]484 ...
485
486 ;; As the pointer goes out of scope, store a null value into
487 ;; it, to indicate that the value is no longer live.
488 store %Object* null, %Object** %X
489 ...
490</pre>
491
492</div>
493
494<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]495<h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]496 <a name="barriers">Reading and writing references in the heap</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]497</h3>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]498
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]499<div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]500
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]501<p>Some collectors need to be informed when the mutator (the program that needs
502garbage collection) either reads a pointer from or writes a pointer to a field
503of a heap object. The code fragments inserted at these points are called
504<em>read barriers</em> and <em>write barriers</em>, respectively. The amount of
505code that needs to be executed is usually quite small and not on the critical
506path of any computation, so the overall performance impact of the barrier is
507tolerable.</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]508
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]509<p>Barriers often require access to the <em>object pointer</em> rather than the
510<em>derived pointer</em> (which is a pointer to the field within the
511object). Accordingly, these intrinsics take both pointers as separate arguments
512for completeness. In this snippet, <tt>%object</tt> is the object pointer, and
513<tt>%derived</tt> is the derived pointer:</p>
514
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]515<blockquote><pre>
516 ;; An array type.
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]517 %class.Array = type { %class.Object, i32, [0 x %class.Object*] }
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]518 ...
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]519
520 ;; Load the object pointer from a gcroot.
521 %object = load %class.Array** %object_addr
522
523 ;; Compute the derived pointer.
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]524 %derived = getelementptr %object, i32 0, i32 2, i32 %n</pre></blockquote>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]525
Gordon Henriksenbffc8662009-03-06 02:42:47 +0000[diff] [blame]526<p>LLVM does not enforce this relationship between the object and derived
527pointer (although a <a href="#plugin">plugin</a> might). However, it would be
528an unusual collector that violated it.</p>
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]529
Gordon Henriksenbffc8662009-03-06 02:42:47 +0000[diff] [blame]530<p>The use of these intrinsics is naturally optional if the target GC does
531require the corresponding barrier. Such a GC plugin will replace the intrinsic
532calls with the corresponding <tt>load</tt> or <tt>store</tt> instruction if they
533are used.</p>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]534
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]535<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]536<h4>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]537 <a name="gcwrite">Write barrier: <tt>llvm.gcwrite</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]538</h4>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]539
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]540<div class="doc_code"><tt>
541void @llvm.gcwrite(i8* %value, i8* %object, i8** %derived)
542</tt></div>
543
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]544<div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]545
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]546<p>For write barriers, LLVM provides the <tt>llvm.gcwrite</tt> intrinsic
547function. It has exactly the same semantics as a non-volatile <tt>store</tt> to
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]548the derived pointer (the third argument). The exact code generated is specified
549by a <a href="#plugin">compiler plugin</a>.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]550
551<p>Many important algorithms require write barriers, including generational
552and concurrent collectors. Additionally, write barriers could be used to
553implement reference counting.</p>
554
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]555</div>
556
557<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]558<h4>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]559 <a name="gcread">Read barrier: <tt>llvm.gcread</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]560</h4>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]561
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]562<div class="doc_code"><tt>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]563i8* @llvm.gcread(i8* %object, i8** %derived)<br>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]564</tt></div>
565
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]566<div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]567
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]568<p>For read barriers, LLVM provides the <tt>llvm.gcread</tt> intrinsic function.
569It has exactly the same semantics as a non-volatile <tt>load</tt> from the
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]570derived pointer (the second argument). The exact code generated is specified by
571a <a href="#plugin">compiler plugin</a>.</p>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]572
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]573<p>Read barriers are needed by fewer algorithms than write barriers, and may
574have a greater performance impact since pointer reads are more frequent than
575writes.</p>
576
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]577</div>
578
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]579</div>
580
581</div>
582
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]583<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]584<h2>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]585 <a name="plugin">Implementing a collector plugin</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]586</h2>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]587<!-- *********************************************************************** -->
588
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]589<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]590
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]591<p>User code specifies which GC code generation to use with the <tt>gc</tt>
592function attribute or, equivalently, with the <tt>setGC</tt> method of
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]593<tt>Function</tt>.</p>
594
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]595<p>To implement a GC plugin, it is necessary to subclass
596<tt>llvm::GCStrategy</tt>, which can be accomplished in a few lines of
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]597boilerplate code. LLVM's infrastructure provides access to several important
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]598algorithms. For an uncontroversial collector, all that remains may be to
599compile LLVM's computed stack map to assembly code (using the binary
600representation expected by the runtime library). This can be accomplished in
601about 100 lines of code.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]602
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]603<p>This is not the appropriate place to implement a garbage collected heap or a
604garbage collector itself. That code should exist in the language's runtime
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]605library. The compiler plugin is responsible for generating code which
606conforms to the binary interface defined by library, most essentially the
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]607<a href="#stack-map">stack map</a>.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]608
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]609<p>To subclass <tt>llvm::GCStrategy</tt> and register it with the compiler:</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]610
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]611<blockquote><pre>// lib/MyGC/MyGC.cpp - Example LLVM GC plugin
612
613#include "llvm/CodeGen/GCStrategy.h"
614#include "llvm/CodeGen/GCMetadata.h"
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]615#include "llvm/Support/Compiler.h"
616
617using namespace llvm;
618
619namespace {
Duncan Sands16d8f8b2010-05-11 20:16:09 +0000[diff] [blame]620 class LLVM_LIBRARY_VISIBILITY MyGC : public GCStrategy {
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]621 public:
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]622 MyGC() {}
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]623 };
624
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]625 GCRegistry::Add&lt;MyGC&gt;
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]626 X("mygc", "My bespoke garbage collector.");
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]627}</pre></blockquote>
628
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]629<p>This boilerplate collector does nothing. More specifically:</p>
630
631<ul>
632 <li><tt>llvm.gcread</tt> calls are replaced with the corresponding
633 <tt>load</tt> instruction.</li>
634 <li><tt>llvm.gcwrite</tt> calls are replaced with the corresponding
635 <tt>store</tt> instruction.</li>
636 <li>No safe points are added to the code.</li>
637 <li>The stack map is not compiled into the executable.</li>
638</ul>
639
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]640<p>Using the LLVM makefiles (like the <a
641href="http://llvm.org/viewvc/llvm-project/llvm/trunk/projects/sample/">sample
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]642project</a>), this code can be compiled as a plugin using a simple
643makefile:</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]644
645<blockquote><pre
646># lib/MyGC/Makefile
647
648LEVEL := ../..
649LIBRARYNAME = <var>MyGC</var>
650LOADABLE_MODULE = 1
651
652include $(LEVEL)/Makefile.common</pre></blockquote>
653
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]654<p>Once the plugin is compiled, code using it may be compiled using <tt>llc
655-load=<var>MyGC.so</var></tt> (though <var>MyGC.so</var> may have some other
656platform-specific extension):</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]657
658<blockquote><pre
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]659>$ cat sample.ll
660define void @f() gc "mygc" {
661entry:
662 ret void
663}
664$ llvm-as &lt; sample.ll | llc -load=MyGC.so</pre></blockquote>
665
666<p>It is also possible to statically link the collector plugin into tools, such
667as a language-specific compiler front-end.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]668
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]669<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]670<h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]671 <a name="collector-algos">Overview of available features</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]672</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]673
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]674<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]675
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]676<p><tt>GCStrategy</tt> provides a range of features through which a plugin
677may do useful work. Some of these are callbacks, some are algorithms that can
678be enabled, disabled, or customized. This matrix summarizes the supported (and
679planned) features and correlates them with the collection techniques which
680typically require them.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]681
682<table>
683 <tr>
684 <th>Algorithm</th>
685 <th>Done</th>
686 <th>shadow stack</th>
687 <th>refcount</th>
688 <th>mark-sweep</th>
689 <th>copying</th>
690 <th>incremental</th>
691 <th>threaded</th>
692 <th>concurrent</th>
693 </tr>
694 <tr>
695 <th class="rowhead"><a href="#stack-map">stack map</a></th>
696 <td>&#10004;</td>
697 <td></td>
698 <td></td>
699 <td>&#10008;</td>
700 <td>&#10008;</td>
701 <td>&#10008;</td>
702 <td>&#10008;</td>
703 <td>&#10008;</td>
704 </tr>
705 <tr>
706 <th class="rowhead"><a href="#init-roots">initialize roots</a></th>
707 <td>&#10004;</td>
708 <td>&#10008;</td>
709 <td>&#10008;</td>
710 <td>&#10008;</td>
711 <td>&#10008;</td>
712 <td>&#10008;</td>
713 <td>&#10008;</td>
714 <td>&#10008;</td>
715 </tr>
716 <tr class="doc_warning">
717 <th class="rowhead">derived pointers</th>
718 <td>NO</td>
719 <td></td>
720 <td></td>
721 <td></td>
722 <td></td>
723 <td></td>
724 <td>&#10008;*</td>
725 <td>&#10008;*</td>
726 </tr>
727 <tr>
728 <th class="rowhead"><em><a href="#custom">custom lowering</a></em></th>
729 <td>&#10004;</td>
730 <th></th>
731 <th></th>
732 <th></th>
733 <th></th>
734 <th></th>
735 <th></th>
736 <th></th>
737 </tr>
738 <tr>
739 <th class="rowhead indent">gcroot</th>
740 <td>&#10004;</td>
741 <td>&#10008;</td>
742 <td>&#10008;</td>
743 <td></td>
744 <td></td>
745 <td></td>
746 <td></td>
747 <td></td>
748 </tr>
749 <tr>
750 <th class="rowhead indent">gcwrite</th>
751 <td>&#10004;</td>
752 <td></td>
753 <td>&#10008;</td>
754 <td></td>
755 <td></td>
756 <td>&#10008;</td>
757 <td></td>
758 <td>&#10008;</td>
759 </tr>
760 <tr>
761 <th class="rowhead indent">gcread</th>
762 <td>&#10004;</td>
763 <td></td>
764 <td></td>
765 <td></td>
766 <td></td>
767 <td></td>
768 <td></td>
769 <td>&#10008;</td>
770 </tr>
771 <tr>
772 <th class="rowhead"><em><a href="#safe-points">safe points</a></em></th>
773 <td></td>
774 <th></th>
775 <th></th>
776 <th></th>
777 <th></th>
778 <th></th>
779 <th></th>
780 <th></th>
781 </tr>
782 <tr>
783 <th class="rowhead indent">in calls</th>
784 <td>&#10004;</td>
785 <td></td>
786 <td></td>
787 <td>&#10008;</td>
788 <td>&#10008;</td>
789 <td>&#10008;</td>
790 <td>&#10008;</td>
791 <td>&#10008;</td>
792 </tr>
793 <tr>
794 <th class="rowhead indent">before calls</th>
795 <td>&#10004;</td>
796 <td></td>
797 <td></td>
798 <td></td>
799 <td></td>
800 <td></td>
801 <td>&#10008;</td>
802 <td>&#10008;</td>
803 </tr>
804 <tr class="doc_warning">
805 <th class="rowhead indent">for loops</th>
806 <td>NO</td>
807 <td></td>
808 <td></td>
809 <td></td>
810 <td></td>
811 <td></td>
812 <td>&#10008;</td>
813 <td>&#10008;</td>
814 </tr>
815 <tr>
816 <th class="rowhead indent">before escape</th>
817 <td>&#10004;</td>
818 <td></td>
819 <td></td>
820 <td></td>
821 <td></td>
822 <td></td>
823 <td>&#10008;</td>
824 <td>&#10008;</td>
825 </tr>
826 <tr class="doc_warning">
827 <th class="rowhead">emit code at safe points</th>
828 <td>NO</td>
829 <td></td>
830 <td></td>
831 <td></td>
832 <td></td>
833 <td></td>
834 <td>&#10008;</td>
835 <td>&#10008;</td>
836 </tr>
837 <tr>
838 <th class="rowhead"><em>output</em></th>
839 <td></td>
840 <th></th>
841 <th></th>
842 <th></th>
843 <th></th>
844 <th></th>
845 <th></th>
846 <th></th>
847 </tr>
848 <tr>
849 <th class="rowhead indent"><a href="#assembly">assembly</a></th>
850 <td>&#10004;</td>
851 <td></td>
852 <td></td>
853 <td>&#10008;</td>
854 <td>&#10008;</td>
855 <td>&#10008;</td>
856 <td>&#10008;</td>
857 <td>&#10008;</td>
858 </tr>
859 <tr class="doc_warning">
860 <th class="rowhead indent">JIT</th>
861 <td>NO</td>
862 <td></td>
863 <td></td>
864 <td class="optl">&#10008;</td>
865 <td class="optl">&#10008;</td>
866 <td class="optl">&#10008;</td>
867 <td class="optl">&#10008;</td>
868 <td class="optl">&#10008;</td>
869 </tr>
870 <tr class="doc_warning">
871 <th class="rowhead indent">obj</th>
872 <td>NO</td>
873 <td></td>
874 <td></td>
875 <td class="optl">&#10008;</td>
876 <td class="optl">&#10008;</td>
877 <td class="optl">&#10008;</td>
878 <td class="optl">&#10008;</td>
879 <td class="optl">&#10008;</td>
880 </tr>
881 <tr class="doc_warning">
882 <th class="rowhead">live analysis</th>
883 <td>NO</td>
884 <td></td>
885 <td></td>
886 <td class="optl">&#10008;</td>
887 <td class="optl">&#10008;</td>
888 <td class="optl">&#10008;</td>
889 <td class="optl">&#10008;</td>
890 <td class="optl">&#10008;</td>
891 </tr>
892 <tr class="doc_warning">
893 <th class="rowhead">register map</th>
894 <td>NO</td>
895 <td></td>
896 <td></td>
897 <td class="optl">&#10008;</td>
898 <td class="optl">&#10008;</td>
899 <td class="optl">&#10008;</td>
900 <td class="optl">&#10008;</td>
901 <td class="optl">&#10008;</td>
902 </tr>
903 <tr>
904 <td colspan="10">
905 <div><span class="doc_warning">*</span> Derived pointers only pose a
906 hazard to copying collectors.</div>
907 <div><span class="optl">&#10008;</span> in gray denotes a feature which
908 could be utilized if available.</div>
909 </td>
910 </tr>
911</table>
912
913<p>To be clear, the collection techniques above are defined as:</p>
914
915<dl>
916 <dt>Shadow Stack</dt>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]917 <dd>The mutator carefully maintains a linked list of stack roots.</dd>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]918 <dt>Reference Counting</dt>
919 <dd>The mutator maintains a reference count for each object and frees an
920 object when its count falls to zero.</dd>
921 <dt>Mark-Sweep</dt>
922 <dd>When the heap is exhausted, the collector marks reachable objects starting
923 from the roots, then deallocates unreachable objects in a sweep
924 phase.</dd>
925 <dt>Copying</dt>
926 <dd>As reachability analysis proceeds, the collector copies objects from one
927 heap area to another, compacting them in the process. Copying collectors
928 enable highly efficient "bump pointer" allocation and can improve locality
929 of reference.</dd>
930 <dt>Incremental</dt>
931 <dd>(Including generational collectors.) Incremental collectors generally have
932 all the properties of a copying collector (regardless of whether the
933 mature heap is compacting), but bring the added complexity of requiring
934 write barriers.</dd>
935 <dt>Threaded</dt>
936 <dd>Denotes a multithreaded mutator; the collector must still stop the mutator
937 ("stop the world") before beginning reachability analysis. Stopping a
938 multithreaded mutator is a complicated problem. It generally requires
939 highly platform specific code in the runtime, and the production of
940 carefully designed machine code at safe points.</dd>
941 <dt>Concurrent</dt>
942 <dd>In this technique, the mutator and the collector run concurrently, with
943 the goal of eliminating pause times. In a <em>cooperative</em> collector,
944 the mutator further aids with collection should a pause occur, allowing
945 collection to take advantage of multiprocessor hosts. The "stop the world"
946 problem of threaded collectors is generally still present to a limited
947 extent. Sophisticated marking algorithms are necessary. Read barriers may
948 be necessary.</dd>
949</dl>
950
951<p>As the matrix indicates, LLVM's garbage collection infrastructure is already
952suitable for a wide variety of collectors, but does not currently extend to
953multithreaded programs. This will be added in the future as there is
954interest.</p>
955
956</div>
957
958<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]959<h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]960 <a name="stack-map">Computing stack maps</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]961</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]962
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]963<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]964
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]965<p>LLVM automatically computes a stack map. One of the most important features
966of a <tt>GCStrategy</tt> is to compile this information into the executable in
967the binary representation expected by the runtime library.</p>
968
969<p>The stack map consists of the location and identity of each GC root in the
970each function in the module. For each root:</p>
971
972<ul>
973 <li><tt>RootNum</tt>: The index of the root.</li>
974 <li><tt>StackOffset</tt>: The offset of the object relative to the frame
975 pointer.</li>
976 <li><tt>RootMetadata</tt>: The value passed as the <tt>%metadata</tt>
977 parameter to the <a href="#gcroot"><tt>@llvm.gcroot</tt></a> intrinsic.</li>
978</ul>
979
980<p>Also, for the function as a whole:</p>
981
982<ul>
983 <li><tt>getFrameSize()</tt>: The overall size of the function's initial
984 stack frame, not accounting for any dynamic allocation.</li>
985 <li><tt>roots_size()</tt>: The count of roots in the function.</li>
986</ul>
987
988<p>To access the stack map, use <tt>GCFunctionMetadata::roots_begin()</tt> and
989-<tt>end()</tt> from the <tt><a
990href="#assembly">GCMetadataPrinter</a></tt>:</p>
991
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]992<blockquote><pre
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]993>for (iterator I = begin(), E = end(); I != E; ++I) {
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]994 GCFunctionInfo *FI = *I;
995 unsigned FrameSize = FI-&gt;getFrameSize();
996 size_t RootCount = FI-&gt;roots_size();
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]997
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]998 for (GCFunctionInfo::roots_iterator RI = FI-&gt;roots_begin(),
999 RE = FI-&gt;roots_end();
1000 RI != RE; ++RI) {
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1001 int RootNum = RI->Num;
1002 int RootStackOffset = RI->StackOffset;
1003 Constant *RootMetadata = RI->Metadata;
1004 }
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1005}</pre></blockquote>
1006
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]1007<p>If the <tt>llvm.gcroot</tt> intrinsic is eliminated before code generation by
1008a custom lowering pass, LLVM will compute an empty stack map. This may be useful
1009for collector plugins which implement reference counting or a shadow stack.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1010
1011</div>
1012
1013
1014<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1015<h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1016 <a name="init-roots">Initializing roots to null: <tt>InitRoots</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1017</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1018
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1019<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1020
1021<blockquote><pre
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1022>MyGC::MyGC() {
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1023 InitRoots = true;
1024}</pre></blockquote>
1025
1026<p>When set, LLVM will automatically initialize each root to <tt>null</tt> upon
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1027entry to the function. This prevents the GC's sweep phase from visiting
1028uninitialized pointers, which will almost certainly cause it to crash. This
1029initialization occurs before custom lowering, so the two may be used
1030together.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1031
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1032<p>Since LLVM does not yet compute liveness information, there is no means of
1033distinguishing an uninitialized stack root from an initialized one. Therefore,
1034this feature should be used by all GC plugins. It is enabled by default.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1035
1036</div>
1037
1038
1039<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1040<h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1041 <a name="custom">Custom lowering of intrinsics: <tt>CustomRoots</tt>,
1042 <tt>CustomReadBarriers</tt>, and <tt>CustomWriteBarriers</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1043</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1044
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1045<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1046
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1047<p>For GCs which use barriers or unusual treatment of stack roots, these
1048flags allow the collector to perform arbitrary transformations of the LLVM
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1049IR:</p>
1050
1051<blockquote><pre
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1052>class MyGC : public GCStrategy {
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1053public:
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1054 MyGC() {
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1055 CustomRoots = true;
1056 CustomReadBarriers = true;
1057 CustomWriteBarriers = true;
1058 }
1059
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1060 virtual bool initializeCustomLowering(Module &amp;M);
1061 virtual bool performCustomLowering(Function &amp;F);
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1062};</pre></blockquote>
1063
1064<p>If any of these flags are set, then LLVM suppresses its default lowering for
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1065the corresponding intrinsics and instead calls
1066<tt>performCustomLowering</tt>.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1067
1068<p>LLVM's default action for each intrinsic is as follows:</p>
1069
1070<ul>
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]1071 <li><tt>llvm.gcroot</tt>: Leave it alone. The code generator must see it
1072 or the stack map will not be computed.</li>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1073 <li><tt>llvm.gcread</tt>: Substitute a <tt>load</tt> instruction.</li>
1074 <li><tt>llvm.gcwrite</tt>: Substitute a <tt>store</tt> instruction.</li>
1075</ul>
1076
1077<p>If <tt>CustomReadBarriers</tt> or <tt>CustomWriteBarriers</tt> are specified,
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1078then <tt>performCustomLowering</tt> <strong>must</strong> eliminate the
1079corresponding barriers.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1080
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1081<p><tt>performCustomLowering</tt> must comply with the same restrictions as <a
1082href="WritingAnLLVMPass.html#runOnFunction"><tt
1083>FunctionPass::runOnFunction</tt></a>.
1084Likewise, <tt>initializeCustomLowering</tt> has the same semantics as <a
1085href="WritingAnLLVMPass.html#doInitialization_mod"><tt
1086>Pass::doInitialization(Module&amp;)</tt></a>.</p>
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1087
1088<p>The following can be used as a template:</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1089
1090<blockquote><pre
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1091>#include "llvm/Module.h"
Gordon Henriksen0adede02007-12-22 23:32:32 +0000[diff] [blame]1092#include "llvm/IntrinsicInst.h"
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1093
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1094bool MyGC::initializeCustomLowering(Module &amp;M) {
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1095 return false;
1096}
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1097
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1098bool MyGC::performCustomLowering(Function &amp;F) {
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1099 bool MadeChange = false;
1100
1101 for (Function::iterator BB = F.begin(), E = F.end(); BB != E; ++BB)
Gordon Henriksen74f4ded2007-12-22 23:34:26 +0000[diff] [blame]1102 for (BasicBlock::iterator II = BB-&gt;begin(), E = BB-&gt;end(); II != E; )
1103 if (IntrinsicInst *CI = dyn_cast&lt;IntrinsicInst&gt;(II++))
Gordon Henriksen0adede02007-12-22 23:32:32 +0000[diff] [blame]1104 if (Function *F = CI-&gt;getCalledFunction())
1105 switch (F-&gt;getIntrinsicID()) {
1106 case Intrinsic::gcwrite:
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1107 // Handle llvm.gcwrite.
Gordon Henriksen0adede02007-12-22 23:32:32 +0000[diff] [blame]1108 CI-&gt;eraseFromParent();
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1109 MadeChange = true;
Gordon Henriksen0adede02007-12-22 23:32:32 +0000[diff] [blame]1110 break;
1111 case Intrinsic::gcread:
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1112 // Handle llvm.gcread.
Gordon Henriksen0adede02007-12-22 23:32:32 +0000[diff] [blame]1113 CI-&gt;eraseFromParent();
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1114 MadeChange = true;
Gordon Henriksen0adede02007-12-22 23:32:32 +0000[diff] [blame]1115 break;
1116 case Intrinsic::gcroot:
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1117 // Handle llvm.gcroot.
Gordon Henriksen0adede02007-12-22 23:32:32 +0000[diff] [blame]1118 CI-&gt;eraseFromParent();
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1119 MadeChange = true;
Gordon Henriksen0adede02007-12-22 23:32:32 +0000[diff] [blame]1120 break;
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1121 }
1122
1123 return MadeChange;
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1124}</pre></blockquote>
1125
1126</div>
1127
1128
1129<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1130<h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1131 <a name="safe-points">Generating safe points: <tt>NeededSafePoints</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1132</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1133
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1134<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1135
1136<p>LLVM can compute four kinds of safe points:</p>
1137
1138<blockquote><pre
1139>namespace GC {
1140 /// PointKind - The type of a collector-safe point.
1141 ///
1142 enum PointKind {
1143 Loop, //&lt; Instr is a loop (backwards branch).
1144 Return, //&lt; Instr is a return instruction.
1145 PreCall, //&lt; Instr is a call instruction.
1146 PostCall //&lt; Instr is the return address of a call.
1147 };
1148}</pre></blockquote>
1149
1150<p>A collector can request any combination of the four by setting the
1151<tt>NeededSafePoints</tt> mask:</p>
1152
1153<blockquote><pre
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1154>MyGC::MyGC() {
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1155 NeededSafePoints = 1 &lt;&lt; GC::Loop
1156 | 1 &lt;&lt; GC::Return
1157 | 1 &lt;&lt; GC::PreCall
1158 | 1 &lt;&lt; GC::PostCall;
1159}</pre></blockquote>
1160
1161<p>It can then use the following routines to access safe points.</p>
1162
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1163<blockquote><pre
1164>for (iterator I = begin(), E = end(); I != E; ++I) {
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1165 GCFunctionInfo *MD = *I;
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1166 size_t PointCount = MD-&gt;size();
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1167
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1168 for (GCFunctionInfo::iterator PI = MD-&gt;begin(),
1169 PE = MD-&gt;end(); PI != PE; ++PI) {
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1170 GC::PointKind PointKind = PI-&gt;Kind;
1171 unsigned PointNum = PI-&gt;Num;
1172 }
1173}
1174</pre></blockquote>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1175
1176<p>Almost every collector requires <tt>PostCall</tt> safe points, since these
1177correspond to the moments when the function is suspended during a call to a
1178subroutine.</p>
1179
1180<p>Threaded programs generally require <tt>Loop</tt> safe points to guarantee
1181that the application will reach a safe point within a bounded amount of time,
1182even if it is executing a long-running loop which contains no function
1183calls.</p>
1184
1185<p>Threaded collectors may also require <tt>Return</tt> and <tt>PreCall</tt>
1186safe points to implement "stop the world" techniques using self-modifying code,
1187where it is important that the program not exit the function without reaching a
1188safe point (because only the topmost function has been patched).</p>
1189
1190</div>
1191
1192
1193<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1194<h3>
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1195 <a name="assembly">Emitting assembly code: <tt>GCMetadataPrinter</tt></a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1196</h3>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1197
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1198<div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1199
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]1200<p>LLVM allows a plugin to print arbitrary assembly code before and after the
1201rest of a module's assembly code. At the end of the module, the GC can compile
1202the LLVM stack map into assembly code. (At the beginning, this information is not
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1203yet computed.)</p>
1204
1205<p>Since AsmWriter and CodeGen are separate components of LLVM, a separate
1206abstract base class and registry is provided for printing assembly code, the
Gordon Henriksenc9c0b592009-03-02 03:47:20 +0000[diff] [blame]1207<tt>GCMetadaPrinter</tt> and <tt>GCMetadataPrinterRegistry</tt>. The AsmWriter
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1208will look for such a subclass if the <tt>GCStrategy</tt> sets
1209<tt>UsesMetadata</tt>:</p>
1210
1211<blockquote><pre
1212>MyGC::MyGC() {
1213 UsesMetadata = true;
1214}</pre></blockquote>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1215
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]1216<p>This separation allows JIT-only clients to be smaller.</p>
1217
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1218<p>Note that LLVM does not currently have analogous APIs to support code
1219generation in the JIT, nor using the object writers.</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1220
1221<blockquote><pre
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1222>// lib/MyGC/MyGCPrinter.cpp - Example LLVM GC printer
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1223
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1224#include "llvm/CodeGen/GCMetadataPrinter.h"
1225#include "llvm/Support/Compiler.h"
1226
1227using namespace llvm;
1228
1229namespace {
Duncan Sands16d8f8b2010-05-11 20:16:09 +0000[diff] [blame]1230 class LLVM_LIBRARY_VISIBILITY MyGCPrinter : public GCMetadataPrinter {
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1231 public:
1232 virtual void beginAssembly(std::ostream &amp;OS, AsmPrinter &amp;AP,
1233 const TargetAsmInfo &amp;TAI);
1234
1235 virtual void finishAssembly(std::ostream &amp;OS, AsmPrinter &amp;AP,
1236 const TargetAsmInfo &amp;TAI);
1237 };
1238
1239 GCMetadataPrinterRegistry::Add&lt;MyGCPrinter&gt;
1240 X("mygc", "My bespoke garbage collector.");
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1241}</pre></blockquote>
1242
1243<p>The collector should use <tt>AsmPrinter</tt> and <tt>TargetAsmInfo</tt> to
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1244print portable assembly code to the <tt>std::ostream</tt>. The collector itself
1245contains the stack map for the entire module, and may access the
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1246<tt>GCFunctionInfo</tt> using its own <tt>begin()</tt> and <tt>end()</tt>
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1247methods. Here's a realistic example:</p>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1248
1249<blockquote><pre
1250>#include "llvm/CodeGen/AsmPrinter.h"
1251#include "llvm/Function.h"
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1252#include "llvm/Target/TargetMachine.h"
1253#include "llvm/Target/TargetData.h"
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1254#include "llvm/Target/TargetAsmInfo.h"
1255
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1256void MyGCPrinter::beginAssembly(std::ostream &amp;OS, AsmPrinter &amp;AP,
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1257 const TargetAsmInfo &amp;TAI) {
1258 // Nothing to do.
1259}
1260
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1261void MyGCPrinter::finishAssembly(std::ostream &amp;OS, AsmPrinter &amp;AP,
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1262 const TargetAsmInfo &amp;TAI) {
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1263 // Set up for emitting addresses.
1264 const char *AddressDirective;
1265 int AddressAlignLog;
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1266 if (AP.TM.getTargetData()->getPointerSize() == sizeof(int32_t)) {
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1267 AddressDirective = TAI.getData32bitsDirective();
1268 AddressAlignLog = 2;
1269 } else {
1270 AddressDirective = TAI.getData64bitsDirective();
1271 AddressAlignLog = 3;
1272 }
1273
1274 // Put this in the data section.
1275 AP.SwitchToDataSection(TAI.getDataSection());
1276
1277 // For each function...
Gordon Henriksenad93c4f2007-12-11 00:30:17 +0000[diff] [blame]1278 for (iterator FI = begin(), FE = end(); FI != FE; ++FI) {
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1279 GCFunctionInfo &amp;MD = **FI;
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1280
1281 // Emit this data structure:
1282 //
1283 // struct {
1284 // int32_t PointCount;
1285 // struct {
1286 // void *SafePointAddress;
1287 // int32_t LiveCount;
1288 // int32_t LiveOffsets[LiveCount];
1289 // } Points[PointCount];
1290 // } __gcmap_&lt;FUNCTIONNAME&gt;;
1291
1292 // Align to address width.
1293 AP.EmitAlignment(AddressAlignLog);
1294
Gordon Henriksene5abac42009-03-06 01:57:32 +0000[diff] [blame]1295 // Emit the symbol by which the stack map entry can be found.
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1296 std::string Symbol;
1297 Symbol += TAI.getGlobalPrefix();
1298 Symbol += "__gcmap_";
1299 Symbol += MD.getFunction().getName();
1300 if (const char *GlobalDirective = TAI.getGlobalDirective())
1301 OS &lt;&lt; GlobalDirective &lt;&lt; Symbol &lt;&lt; "\n";
1302 OS &lt;&lt; TAI.getGlobalPrefix() &lt;&lt; Symbol &lt;&lt; ":\n";
1303
1304 // Emit PointCount.
1305 AP.EmitInt32(MD.size());
1306 AP.EOL("safe point count");
1307
1308 // And each safe point...
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1309 for (GCFunctionInfo::iterator PI = MD.begin(),
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1310 PE = MD.end(); PI != PE; ++PI) {
1311 // Align to address width.
1312 AP.EmitAlignment(AddressAlignLog);
1313
1314 // Emit the address of the safe point.
1315 OS &lt;&lt; AddressDirective
1316 &lt;&lt; TAI.getPrivateGlobalPrefix() &lt;&lt; "label" &lt;&lt; PI-&gt;Num;
1317 AP.EOL("safe point address");
1318
1319 // Emit the stack frame size.
1320 AP.EmitInt32(MD.getFrameSize());
1321 AP.EOL("stack frame size");
1322
1323 // Emit the number of live roots in the function.
1324 AP.EmitInt32(MD.live_size(PI));
1325 AP.EOL("live root count");
1326
1327 // And for each live root...
Gordon Henriksen01571ef2008-08-24 03:18:23 +0000[diff] [blame]1328 for (GCFunctionInfo::live_iterator LI = MD.live_begin(PI),
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1329 LE = MD.live_end(PI);
1330 LI != LE; ++LI) {
1331 // Print its offset within the stack frame.
1332 AP.EmitInt32(LI-&gt;StackOffset);
1333 AP.EOL("stack offset");
1334 }
1335 }
1336 }
1337}
1338</pre></blockquote>
1339
1340</div>
1341
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1342</div>
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1343
1344<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1345<h2>
Chris Lattner9b2a1842004-05-27 05:52:10 +0000[diff] [blame]1346 <a name="references">References</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1347</h2>
Chris Lattner9b2a1842004-05-27 05:52:10 +0000[diff] [blame]1348<!-- *********************************************************************** -->
1349
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1350<div>
Chris Lattner9b2a1842004-05-27 05:52:10 +0000[diff] [blame]1351
1352<p><a name="appel89">[Appel89]</a> Runtime Tags Aren't Necessary. Andrew
1353W. Appel. Lisp and Symbolic Computation 19(7):703-705, July 1989.</p>
1354
1355<p><a name="goldberg91">[Goldberg91]</a> Tag-free garbage collection for
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1356strongly typed programming languages. Benjamin Goldberg. ACM SIGPLAN
Chris Lattner9b2a1842004-05-27 05:52:10 +0000[diff] [blame]1357PLDI'91.</p>
1358
1359<p><a name="tolmach94">[Tolmach94]</a> Tag-free garbage collection using
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1360explicit type parameters. Andrew Tolmach. Proceedings of the 1994 ACM
Chris Lattner9b2a1842004-05-27 05:52:10 +0000[diff] [blame]1361conference on LISP and functional programming.</p>
1362
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1363<p><a name="henderson02">[Henderson2002]</a> <a
1364href="http://citeseer.ist.psu.edu/henderson02accurate.html">
1365Accurate Garbage Collection in an Uncooperative Environment</a>.
1366Fergus Henderson. International Symposium on Memory Management 2002.</p>
1367
Chris Lattner9b2a1842004-05-27 05:52:10 +0000[diff] [blame]1368</div>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]1369
Gordon Henriksen326e24f2007-09-27 19:31:36 +0000[diff] [blame]1370
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]1371<!-- *********************************************************************** -->
1372
1373<hr>
1374<address>
1375 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]1376 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]1377 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]1378 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]1379
1380 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]1381 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
Chris Lattner0d8c2db2004-05-23 21:02:20 +0000[diff] [blame]1382 Last modified: $Date$
1383</address>
1384
1385</body>
1386</html>