1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <meta http-equiv="Content-Type" Content="text/html; charset=UTF-8" >
6 <title>Accurate Garbage Collection with LLVM</title>
7 <link rel="stylesheet" href="llvm.css" type="text/css">
8 <style type="text/css">
9 .rowhead { text-align: left; background: inherit; }
10 .indent { padding-left: 1em; }
11 .optl { color: #BFBFBF; }
17 Accurate Garbage Collection with LLVM
21 <li><a href="#introduction">Introduction</a>
23 <li><a href="#feature">Goals and non-goals</a></li>
27 <li><a href="#quickstart">Getting started</a>
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>
35 <li><a href="#core">Core support</a>
37 <li><a href="#gcattr">Specifying GC code generation:
38 <tt>gc "..."</tt></a></li>
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>
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>
50 <li><a href="#plugin">Compiler plugin interface</a>
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:
61 <tt>GCMetadataPrinter</tt></a></li>
65 <li><a href="#runtime-impl">Implementing a collector runtime</a>
67 <li><a href="#gcdescriptors">Tracing GC pointers from heap
72 <li><a href="#references">References</a></li>
76 <div class="doc_author">
77 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a> and
81 <!-- *********************************************************************** -->
83 <a name="introduction">Introduction</a>
85 <!-- *********************************************************************** -->
89 <p>Garbage collection is a widely used technique that frees the programmer from
90 having to know the lifetimes of heap objects, making software easier to produce
91 and maintain. Many programming languages rely on garbage collection for
92 automatic memory management. There are two primary forms of garbage collection:
93 conservative and accurate.</p>
95 <p>Conservative garbage collection often does not require any special support
96 from either the language or the compiler: it can handle non-type-safe
97 programming languages (such as C/C++) and does not require any special
98 information from the compiler. The
99 <a href="http://www.hpl.hp.com/personal/Hans_Boehm/gc/">Boehm collector</a> is
100 an example of a state-of-the-art conservative collector.</p>
102 <p>Accurate garbage collection requires the ability to identify all pointers in
103 the program at run-time (which requires that the source-language be type-safe in
104 most cases). Identifying pointers at run-time requires compiler support to
105 locate all places that hold live pointer variables at run-time, including the
106 <a href="#gcroot">processor stack and registers</a>.</p>
108 <p>Conservative garbage collection is attractive because it does not require any
109 special compiler support, but it does have problems. In particular, because the
110 conservative garbage collector cannot <i>know</i> that a particular word in the
111 machine is a pointer, it cannot move live objects in the heap (preventing the
112 use of compacting and generational GC algorithms) and it can occasionally suffer
113 from memory leaks due to integer values that happen to point to objects in the
114 program. In addition, some aggressive compiler transformations can break
115 conservative garbage collectors (though these seem rare in practice).</p>
117 <p>Accurate garbage collectors do not suffer from any of these problems, but
118 they can suffer from degraded scalar optimization of the program. In particular,
119 because the runtime must be able to identify and update all pointers active in
120 the program, some optimizations are less effective. In practice, however, the
121 locality and performance benefits of using aggressive garbage collection
122 techniques dominates any low-level losses.</p>
124 <p>This document describes the mechanisms and interfaces provided by LLVM to
125 support accurate garbage collection.</p>
127 <!-- ======================================================================= -->
129 <a name="feature">Goals and non-goals</a>
134 <p>LLVM's intermediate representation provides <a href="#intrinsics">garbage
135 collection intrinsics</a> that offer support for a broad class of
136 collector models. For instance, the intrinsics permit:</p>
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>
148 <p>We hope that the primitive support built into the LLVM IR is sufficient to
149 support a broad class of garbage collected languages including Scheme, ML, Java,
150 C#, Perl, Python, Lua, Ruby, other scripting languages, and more.</p>
152 <p>However, LLVM does not itself provide a garbage collector—this should
153 be part of your language's runtime library. LLVM provides a framework for
154 compile time <a href="#plugin">code generation plugins</a>. The role of these
155 plugins is to generate code and data structures which conforms to the <em>binary
156 interface</em> specified by the <em>runtime library</em>. This is similar to the
157 relationship between LLVM and DWARF debugging info, for example. The
158 difference primarily lies in the lack of an established standard in the domain
159 of garbage collection—thus the plugins.</p>
161 <p>The aspects of the binary interface with which LLVM's GC support is
165 <li>Creation of GC-safe points within code where collection is allowed to
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
173 <li>Emission of read barriers when loading object references. These are
174 useful for interoperating with concurrent collectors.</li>
177 <p>There are additional areas that LLVM does not directly address:</p>
180 <li>Registration of global roots with the runtime.</li>
181 <li>Registration of stack map entries with the runtime.</li>
182 <li>The functions used by the program to allocate memory, trigger a
183 collection, etc.</li>
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
189 <p>In general, LLVM's support for GC does not include features which can be
190 adequately addressed with other features of the IR and does not specify a
191 particular binary interface. On the plus side, this means that you should be
192 able to integrate LLVM with an existing runtime. On the other hand, it leaves
193 a lot of work for the developer of a novel language. However, it's easy to get
194 started quickly and scale up to a more sophisticated implementation as your
195 compiler matures.</p>
201 <!-- *********************************************************************** -->
203 <a name="quickstart">Getting started</a>
205 <!-- *********************************************************************** -->
209 <p>Using a GC with LLVM implies many things, for example:</p>
212 <li>Write a runtime library or find an existing one which implements a GC
214 <li>Implement a memory allocator.</li>
215 <li>Design a binary interface for the stack map, used to identify
216 references within a stack frame on the machine stack.*</li>
217 <li>Implement a stack crawler to discover functions on the call stack.*</li>
218 <li>Implement a registry for global roots.</li>
219 <li>Design a binary interface for type maps, used to identify references
220 within heap objects.</li>
221 <li>Implement a collection routine bringing together all of the above.</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>
232 <li>Generate type maps according to your runtime's binary interface.</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>
237 <li>Compile LLVM's stack map to the binary form expected by the
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>
245 <p>To help with several of these tasks (those indicated with a *), LLVM
246 includes a highly portable, built-in ShadowStack code generator. It is compiled
247 into <tt>llc</tt> and works even with the interpreter and C backends.</p>
249 <!-- ======================================================================= -->
251 <a name="quickstart-compiler">In your compiler</a>
256 <p>To turn the shadow stack on for your functions, first call:</p>
258 <div class="doc_code"><pre
259 >F.setGC("shadow-stack");</pre></div>
261 <p>for each function your compiler emits. Since the shadow stack is built into
262 LLVM, you do not need to load a plugin.</p>
264 <p>Your compiler must also use <tt>@llvm.gcroot</tt> as documented.
265 Don't forget to create a root for each intermediate value that is generated
266 when 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
270 <p>There's no need to use <tt>@llvm.gcread</tt> and <tt>@llvm.gcwrite</tt> over
271 plain <tt>load</tt> and <tt>store</tt> for now. You will need them when
272 switching to a more advanced GC.</p>
276 <!-- ======================================================================= -->
278 <a name="quickstart-runtime">In your runtime</a>
283 <p>The shadow stack doesn't imply a memory allocation algorithm. A semispace
284 collector or building atop <tt>malloc</tt> are great places to start, and can
285 be implemented with very little code.</p>
287 <p>When it comes time to collect, however, your runtime needs to traverse the
288 stack roots, and for this it needs to integrate with the shadow stack. Luckily,
289 doing so is very simple. (This code is heavily commented to help you
290 understand the data structure, but there are only 20 lines of meaningful
293 <pre class="doc_code">
294 /// @brief The map for a single function's stack frame. One of these is
295 /// compiled as constant data into the executable for each function.
297 /// Storage of metadata values is elided if the %metadata parameter to
298 /// @llvm.gcroot is null.
300 int32_t NumRoots; //< Number of roots in stack frame.
301 int32_t NumMeta; //< Number of metadata entries. May be < NumRoots.
302 const void *Meta[0]; //< Metadata for each root.
305 /// @brief A link in the dynamic shadow stack. One of these is embedded in the
306 /// stack frame of each function on the call stack.
308 StackEntry *Next; //< Link to next stack entry (the caller's).
309 const FrameMap *Map; //< Pointer to constant FrameMap.
310 void *Roots[0]; //< Stack roots (in-place array).
313 /// @brief The head of the singly-linked list of StackEntries. Functions push
314 /// and pop onto this in their prologue and epilogue.
316 /// Since there is only a global list, this technique is not threadsafe.
317 StackEntry *llvm_gc_root_chain;
319 /// @brief Calls Visitor(root, meta) for each GC root on the stack.
320 /// root and meta are exactly the values passed to
321 /// <tt>@llvm.gcroot</tt>.
323 /// Visitor could be a function to recursively mark live objects. Or it
324 /// might copy them to another heap or generation.
326 /// @param Visitor A function to invoke for every GC root on the stack.
327 void visitGCRoots(void (*Visitor)(void **Root, const void *Meta)) {
328 for (StackEntry *R = llvm_gc_root_chain; R; R = R->Next) {
331 // For roots [0, NumMeta), the metadata pointer is in the FrameMap.
332 for (unsigned e = R->Map->NumMeta; i != e; ++i)
333 Visitor(&R->Roots[i], R->Map->Meta[i]);
335 // For roots [NumMeta, NumRoots), the metadata pointer is null.
336 for (unsigned e = R->Map->NumRoots; i != e; ++i)
337 Visitor(&R->Roots[i], NULL);
343 <!-- ======================================================================= -->
345 <a name="shadow-stack">About the shadow stack</a>
350 <p>Unlike many GC algorithms which rely on a cooperative code generator to
351 compile stack maps, this algorithm carefully maintains a linked list of stack
352 roots [<a href="#henderson02">Henderson2002</a>]. This so-called "shadow stack"
353 mirrors the machine stack. Maintaining this data structure is slower than using
354 a stack map compiled into the executable as constant data, but has a significant
355 portability advantage because it requires no special support from the target
356 code generator, and does not require tricky platform-specific code to crawl
357 the machine stack.</p>
359 <p>The tradeoff for this simplicity and portability is:</p>
362 <li>High overhead per function call.</li>
363 <li>Not thread-safe.</li>
366 <p>Still, it's an easy way to get started. After your compiler and runtime are
367 up and running, writing a <a href="#plugin">plugin</a> will allow you to take
368 advantage of <a href="#collector-algos">more advanced GC features</a> of LLVM
369 in order to improve performance.</p>
375 <!-- *********************************************************************** -->
377 <a name="core">IR features</a><a name="intrinsics"></a>
379 <!-- *********************************************************************** -->
383 <p>This section describes the garbage collection facilities provided by the
384 <a href="LangRef.html">LLVM intermediate representation</a>. The exact behavior
385 of 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>
388 <p>These facilities are limited to those strictly necessary; they are not
389 intended to be a complete interface to any garbage collector. A program will
390 need to interface with the GC library using the facilities provided by that
393 <!-- ======================================================================= -->
395 <a name="gcattr">Specifying GC code generation: <tt>gc "..."</tt></a>
400 <div class="doc_code"><tt>
401 define <i>ty</i> @<i>name</i>(...) <span style="text-decoration: underline">gc "<i>name</i>"</span> { ...
404 <p>The <tt>gc</tt> function attribute is used to specify the desired GC style
405 to the compiler. Its programmatic equivalent is the <tt>setGC</tt> method of
406 <tt>Function</tt>.</p>
408 <p>Setting <tt>gc "<i>name</i>"</tt> on a function triggers a search for a
409 matching code generation plugin "<i>name</i>"; it is that plugin which defines
410 the exact nature of the code generated to support GC. If none is found, the
411 compiler will raise an error.</p>
413 <p>Specifying the GC style on a per-function basis allows LLVM to link together
414 programs that use different garbage collection algorithms (or none at all).</p>
418 <!-- ======================================================================= -->
420 <a name="gcroot">Identifying GC roots on the stack: <tt>llvm.gcroot</tt></a>
425 <div class="doc_code"><tt>
426 void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
429 <p>The <tt>llvm.gcroot</tt> intrinsic is used to inform LLVM that a stack
430 variable references an object on the heap and is to be tracked for garbage
431 collection. The exact impact on generated code is specified by a <a
432 href="#plugin">compiler plugin</a>.</p>
434 <p>A compiler which uses mem2reg to raise imperative code using <tt>alloca</tt>
435 into SSA form need only add a call to <tt>@llvm.gcroot</tt> for those variables
436 which a pointers into the GC heap.</p>
438 <p>It is also important to mark intermediate values with <tt>llvm.gcroot</tt>.
439 For 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>
442 <p>The first argument <b>must</b> be a value referring to an alloca instruction
443 or a bitcast of an alloca. The second contains a pointer to metadata that
444 should be associated with the pointer, and <b>must</b> be a constant or global
445 value address. If your target collector uses tags, use a null pointer for
448 <p>The <tt>%metadata</tt> argument can be used to avoid requiring heap objects
449 to have 'isa' pointers or tag bits. [<a href="#appel89">Appel89</a>, <a
450 href="#goldberg91">Goldberg91</a>, <a href="#tolmach94">Tolmach94</a>] If
451 specified, its value will be tracked along with the location of the pointer in
454 <p>Consider the following fragment of Java code:</p>
456 <pre class="doc_code">
458 Object X; // A null-initialized reference to an object
463 <p>This block (which may be located in the middle of a function or in a loop
464 nest), could be compiled to this LLVM code:</p>
466 <pre class="doc_code">
468 ;; In the entry block for the function, allocate the
469 ;; stack space for X, which is an LLVM pointer.
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**
475 call void @llvm.gcroot(i8** %X, i8* null)
478 ;; "CodeBlock" is the block corresponding to the start
479 ;; of the scope above.
481 ;; Java null-initializes pointers.
482 store %Object* null, %Object** %X
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
494 <!-- ======================================================================= -->
496 <a name="barriers">Reading and writing references in the heap</a>
501 <p>Some collectors need to be informed when the mutator (the program that needs
502 garbage collection) either reads a pointer from or writes a pointer to a field
503 of 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
505 code that needs to be executed is usually quite small and not on the critical
506 path of any computation, so the overall performance impact of the barrier is
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
511 object). Accordingly, these intrinsics take both pointers as separate arguments
512 for completeness. In this snippet, <tt>%object</tt> is the object pointer, and
513 <tt>%derived</tt> is the derived pointer:</p>
517 %class.Array = type { %class.Object, i32, [0 x %class.Object*] }
520 ;; Load the object pointer from a gcroot.
521 %object = load %class.Array** %object_addr
523 ;; Compute the derived pointer.
524 %derived = getelementptr %object, i32 0, i32 2, i32 %n</pre></blockquote>
526 <p>LLVM does not enforce this relationship between the object and derived
527 pointer (although a <a href="#plugin">plugin</a> might). However, it would be
528 an unusual collector that violated it.</p>
530 <p>The use of these intrinsics is naturally optional if the target GC does
531 require the corresponding barrier. Such a GC plugin will replace the intrinsic
532 calls with the corresponding <tt>load</tt> or <tt>store</tt> instruction if they
535 <!-- ======================================================================= -->
537 <a name="gcwrite">Write barrier: <tt>llvm.gcwrite</tt></a>
542 <div class="doc_code"><tt>
543 void @llvm.gcwrite(i8* %value, i8* %object, i8** %derived)
546 <p>For write barriers, LLVM provides the <tt>llvm.gcwrite</tt> intrinsic
547 function. It has exactly the same semantics as a non-volatile <tt>store</tt> to
548 the derived pointer (the third argument). The exact code generated is specified
549 by a <a href="#plugin">compiler plugin</a>.</p>
551 <p>Many important algorithms require write barriers, including generational
552 and concurrent collectors. Additionally, write barriers could be used to
553 implement reference counting.</p>
557 <!-- ======================================================================= -->
559 <a name="gcread">Read barrier: <tt>llvm.gcread</tt></a>
564 <div class="doc_code"><tt>
565 i8* @llvm.gcread(i8* %object, i8** %derived)<br>
568 <p>For read barriers, LLVM provides the <tt>llvm.gcread</tt> intrinsic function.
569 It has exactly the same semantics as a non-volatile <tt>load</tt> from the
570 derived pointer (the second argument). The exact code generated is specified by
571 a <a href="#plugin">compiler plugin</a>.</p>
573 <p>Read barriers are needed by fewer algorithms than write barriers, and may
574 have a greater performance impact since pointer reads are more frequent than
583 <!-- *********************************************************************** -->
585 <a name="plugin">Implementing a collector plugin</a>
587 <!-- *********************************************************************** -->
591 <p>User code specifies which GC code generation to use with the <tt>gc</tt>
592 function attribute or, equivalently, with the <tt>setGC</tt> method of
593 <tt>Function</tt>.</p>
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
597 boilerplate code. LLVM's infrastructure provides access to several important
598 algorithms. For an uncontroversial collector, all that remains may be to
599 compile LLVM's computed stack map to assembly code (using the binary
600 representation expected by the runtime library). This can be accomplished in
601 about 100 lines of code.</p>
603 <p>This is not the appropriate place to implement a garbage collected heap or a
604 garbage collector itself. That code should exist in the language's runtime
605 library. The compiler plugin is responsible for generating code which
606 conforms to the binary interface defined by library, most essentially the
607 <a href="#stack-map">stack map</a>.</p>
609 <p>To subclass <tt>llvm::GCStrategy</tt> and register it with the compiler:</p>
611 <blockquote><pre>// lib/MyGC/MyGC.cpp - Example LLVM GC plugin
613 #include "llvm/CodeGen/GCStrategy.h"
614 #include "llvm/CodeGen/GCMetadata.h"
615 #include "llvm/Support/Compiler.h"
617 using namespace llvm;
620 class LLVM_LIBRARY_VISIBILITY MyGC : public GCStrategy {
625 GCRegistry::Add<MyGC>
626 X("mygc", "My bespoke garbage collector.");
629 <p>This boilerplate collector does nothing. More specifically:</p>
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>
640 <p>Using the LLVM makefiles (like the <a
641 href="http://llvm.org/viewvc/llvm-project/llvm/trunk/projects/sample/">sample
642 project</a>), this code can be compiled as a plugin using a simple
649 LIBRARYNAME = <var>MyGC</var>
652 include $(LEVEL)/Makefile.common</pre></blockquote>
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
656 platform-specific extension):</p>
660 define void @f() gc "mygc" {
664 $ llvm-as < sample.ll | llc -load=MyGC.so</pre></blockquote>
666 <p>It is also possible to statically link the collector plugin into tools, such
667 as a language-specific compiler front-end.</p>
669 <!-- ======================================================================= -->
671 <a name="collector-algos">Overview of available features</a>
676 <p><tt>GCStrategy</tt> provides a range of features through which a plugin
677 may do useful work. Some of these are callbacks, some are algorithms that can
678 be enabled, disabled, or customized. This matrix summarizes the supported (and
679 planned) features and correlates them with the collection techniques which
680 typically require them.</p>
686 <th>shadow stack</th>
695 <th class="rowhead"><a href="#stack-map">stack map</a></th>
706 <th class="rowhead"><a href="#init-roots">initialize roots</a></th>
716 <tr class="doc_warning">
717 <th class="rowhead">derived pointers</th>
728 <th class="rowhead"><em><a href="#custom">custom lowering</a></em></th>
739 <th class="rowhead indent">gcroot</th>
750 <th class="rowhead indent">gcwrite</th>
761 <th class="rowhead indent">gcread</th>
772 <th class="rowhead"><em><a href="#safe-points">safe points</a></em></th>
783 <th class="rowhead indent">in calls</th>
794 <th class="rowhead indent">before calls</th>
804 <tr class="doc_warning">
805 <th class="rowhead indent">for loops</th>
816 <th class="rowhead indent">before escape</th>
826 <tr class="doc_warning">
827 <th class="rowhead">emit code at safe points</th>
838 <th class="rowhead"><em>output</em></th>
849 <th class="rowhead indent"><a href="#assembly">assembly</a></th>
859 <tr class="doc_warning">
860 <th class="rowhead indent">JIT</th>
864 <td class="optl">✘</td>
865 <td class="optl">✘</td>
866 <td class="optl">✘</td>
867 <td class="optl">✘</td>
868 <td class="optl">✘</td>
870 <tr class="doc_warning">
871 <th class="rowhead indent">obj</th>
875 <td class="optl">✘</td>
876 <td class="optl">✘</td>
877 <td class="optl">✘</td>
878 <td class="optl">✘</td>
879 <td class="optl">✘</td>
881 <tr class="doc_warning">
882 <th class="rowhead">live analysis</th>
886 <td class="optl">✘</td>
887 <td class="optl">✘</td>
888 <td class="optl">✘</td>
889 <td class="optl">✘</td>
890 <td class="optl">✘</td>
892 <tr class="doc_warning">
893 <th class="rowhead">register map</th>
897 <td class="optl">✘</td>
898 <td class="optl">✘</td>
899 <td class="optl">✘</td>
900 <td class="optl">✘</td>
901 <td class="optl">✘</td>
905 <div><span class="doc_warning">*</span> Derived pointers only pose a
906 hazard to copying collectors.</div>
907 <div><span class="optl">✘</span> in gray denotes a feature which
908 could be utilized if available.</div>
913 <p>To be clear, the collection techniques above are defined as:</p>
916 <dt>Shadow Stack</dt>
917 <dd>The mutator carefully maintains a linked list of stack roots.</dd>
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>
922 <dd>When the heap is exhausted, the collector marks reachable objects starting
923 from the roots, then deallocates unreachable objects in a sweep
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
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
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>
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
951 <p>As the matrix indicates, LLVM's garbage collection infrastructure is already
952 suitable for a wide variety of collectors, but does not currently extend to
953 multithreaded programs. This will be added in the future as there is
958 <!-- ======================================================================= -->
960 <a name="stack-map">Computing stack maps</a>
965 <p>LLVM automatically computes a stack map. One of the most important features
966 of a <tt>GCStrategy</tt> is to compile this information into the executable in
967 the binary representation expected by the runtime library.</p>
969 <p>The stack map consists of the location and identity of each GC root in the
970 each function in the module. For each root:</p>
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
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>
980 <p>Also, for the function as a whole:</p>
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>
988 <p>To access the stack map, use <tt>GCFunctionMetadata::roots_begin()</tt> and
989 -<tt>end()</tt> from the <tt><a
990 href="#assembly">GCMetadataPrinter</a></tt>:</p>
993 >for (iterator I = begin(), E = end(); I != E; ++I) {
994 GCFunctionInfo *FI = *I;
995 unsigned FrameSize = FI->getFrameSize();
996 size_t RootCount = FI->roots_size();
998 for (GCFunctionInfo::roots_iterator RI = FI->roots_begin(),
999 RE = FI->roots_end();
1001 int RootNum = RI->Num;
1002 int RootStackOffset = RI->StackOffset;
1003 Constant *RootMetadata = RI->Metadata;
1005 }</pre></blockquote>
1007 <p>If the <tt>llvm.gcroot</tt> intrinsic is eliminated before code generation by
1008 a custom lowering pass, LLVM will compute an empty stack map. This may be useful
1009 for collector plugins which implement reference counting or a shadow stack.</p>
1014 <!-- ======================================================================= -->
1016 <a name="init-roots">Initializing roots to null: <tt>InitRoots</tt></a>
1024 }</pre></blockquote>
1026 <p>When set, LLVM will automatically initialize each root to <tt>null</tt> upon
1027 entry to the function. This prevents the GC's sweep phase from visiting
1028 uninitialized pointers, which will almost certainly cause it to crash. This
1029 initialization occurs before custom lowering, so the two may be used
1032 <p>Since LLVM does not yet compute liveness information, there is no means of
1033 distinguishing an uninitialized stack root from an initialized one. Therefore,
1034 this feature should be used by all GC plugins. It is enabled by default.</p>
1039 <!-- ======================================================================= -->
1041 <a name="custom">Custom lowering of intrinsics: <tt>CustomRoots</tt>,
1042 <tt>CustomReadBarriers</tt>, and <tt>CustomWriteBarriers</tt></a>
1047 <p>For GCs which use barriers or unusual treatment of stack roots, these
1048 flags allow the collector to perform arbitrary transformations of the LLVM
1052 >class MyGC : public GCStrategy {
1056 CustomReadBarriers = true;
1057 CustomWriteBarriers = true;
1060 virtual bool initializeCustomLowering(Module &M);
1061 virtual bool performCustomLowering(Function &F);
1062 };</pre></blockquote>
1064 <p>If any of these flags are set, then LLVM suppresses its default lowering for
1065 the corresponding intrinsics and instead calls
1066 <tt>performCustomLowering</tt>.</p>
1068 <p>LLVM's default action for each intrinsic is as follows:</p>
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>
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>
1077 <p>If <tt>CustomReadBarriers</tt> or <tt>CustomWriteBarriers</tt> are specified,
1078 then <tt>performCustomLowering</tt> <strong>must</strong> eliminate the
1079 corresponding barriers.</p>
1081 <p><tt>performCustomLowering</tt> must comply with the same restrictions as <a
1082 href="WritingAnLLVMPass.html#runOnFunction"><tt
1083 >FunctionPass::runOnFunction</tt></a>.
1084 Likewise, <tt>initializeCustomLowering</tt> has the same semantics as <a
1085 href="WritingAnLLVMPass.html#doInitialization_mod"><tt
1086 >Pass::doInitialization(Module&)</tt></a>.</p>
1088 <p>The following can be used as a template:</p>
1091 >#include "llvm/Module.h"
1092 #include "llvm/IntrinsicInst.h"
1094 bool MyGC::initializeCustomLowering(Module &M) {
1098 bool MyGC::performCustomLowering(Function &F) {
1099 bool MadeChange = false;
1101 for (Function::iterator BB = F.begin(), E = F.end(); BB != E; ++BB)
1102 for (BasicBlock::iterator II = BB->begin(), E = BB->end(); II != E; )
1103 if (IntrinsicInst *CI = dyn_cast<IntrinsicInst>(II++))
1104 if (Function *F = CI->getCalledFunction())
1105 switch (F->getIntrinsicID()) {
1106 case Intrinsic::gcwrite:
1107 // Handle llvm.gcwrite.
1108 CI->eraseFromParent();
1111 case Intrinsic::gcread:
1112 // Handle llvm.gcread.
1113 CI->eraseFromParent();
1116 case Intrinsic::gcroot:
1117 // Handle llvm.gcroot.
1118 CI->eraseFromParent();
1124 }</pre></blockquote>
1129 <!-- ======================================================================= -->
1131 <a name="safe-points">Generating safe points: <tt>NeededSafePoints</tt></a>
1136 <p>LLVM can compute four kinds of safe points:</p>
1140 /// PointKind - The type of a collector-safe point.
1143 Loop, //< Instr is a loop (backwards branch).
1144 Return, //< Instr is a return instruction.
1145 PreCall, //< Instr is a call instruction.
1146 PostCall //< Instr is the return address of a call.
1148 }</pre></blockquote>
1150 <p>A collector can request any combination of the four by setting the
1151 <tt>NeededSafePoints</tt> mask:</p>
1155 NeededSafePoints = 1 << GC::Loop
1156 | 1 << GC::Return
1157 | 1 << GC::PreCall
1158 | 1 << GC::PostCall;
1159 }</pre></blockquote>
1161 <p>It can then use the following routines to access safe points.</p>
1164 >for (iterator I = begin(), E = end(); I != E; ++I) {
1165 GCFunctionInfo *MD = *I;
1166 size_t PointCount = MD->size();
1168 for (GCFunctionInfo::iterator PI = MD->begin(),
1169 PE = MD->end(); PI != PE; ++PI) {
1170 GC::PointKind PointKind = PI->Kind;
1171 unsigned PointNum = PI->Num;
1176 <p>Almost every collector requires <tt>PostCall</tt> safe points, since these
1177 correspond to the moments when the function is suspended during a call to a
1180 <p>Threaded programs generally require <tt>Loop</tt> safe points to guarantee
1181 that the application will reach a safe point within a bounded amount of time,
1182 even if it is executing a long-running loop which contains no function
1185 <p>Threaded collectors may also require <tt>Return</tt> and <tt>PreCall</tt>
1186 safe points to implement "stop the world" techniques using self-modifying code,
1187 where it is important that the program not exit the function without reaching a
1188 safe point (because only the topmost function has been patched).</p>
1193 <!-- ======================================================================= -->
1195 <a name="assembly">Emitting assembly code: <tt>GCMetadataPrinter</tt></a>
1200 <p>LLVM allows a plugin to print arbitrary assembly code before and after the
1201 rest of a module's assembly code. At the end of the module, the GC can compile
1202 the LLVM stack map into assembly code. (At the beginning, this information is not
1205 <p>Since AsmWriter and CodeGen are separate components of LLVM, a separate
1206 abstract base class and registry is provided for printing assembly code, the
1207 <tt>GCMetadaPrinter</tt> and <tt>GCMetadataPrinterRegistry</tt>. The AsmWriter
1208 will look for such a subclass if the <tt>GCStrategy</tt> sets
1209 <tt>UsesMetadata</tt>:</p>
1213 UsesMetadata = true;
1214 }</pre></blockquote>
1216 <p>This separation allows JIT-only clients to be smaller.</p>
1218 <p>Note that LLVM does not currently have analogous APIs to support code
1219 generation in the JIT, nor using the object writers.</p>
1222 >// lib/MyGC/MyGCPrinter.cpp - Example LLVM GC printer
1224 #include "llvm/CodeGen/GCMetadataPrinter.h"
1225 #include "llvm/Support/Compiler.h"
1227 using namespace llvm;
1230 class LLVM_LIBRARY_VISIBILITY MyGCPrinter : public GCMetadataPrinter {
1232 virtual void beginAssembly(std::ostream &OS, AsmPrinter &AP,
1233 const TargetAsmInfo &TAI);
1235 virtual void finishAssembly(std::ostream &OS, AsmPrinter &AP,
1236 const TargetAsmInfo &TAI);
1239 GCMetadataPrinterRegistry::Add<MyGCPrinter>
1240 X("mygc", "My bespoke garbage collector.");
1241 }</pre></blockquote>
1243 <p>The collector should use <tt>AsmPrinter</tt> and <tt>TargetAsmInfo</tt> to
1244 print portable assembly code to the <tt>std::ostream</tt>. The collector itself
1245 contains the stack map for the entire module, and may access the
1246 <tt>GCFunctionInfo</tt> using its own <tt>begin()</tt> and <tt>end()</tt>
1247 methods. Here's a realistic example:</p>
1250 >#include "llvm/CodeGen/AsmPrinter.h"
1251 #include "llvm/Function.h"
1252 #include "llvm/Target/TargetMachine.h"
1253 #include "llvm/Target/TargetData.h"
1254 #include "llvm/Target/TargetAsmInfo.h"
1256 void MyGCPrinter::beginAssembly(std::ostream &OS, AsmPrinter &AP,
1257 const TargetAsmInfo &TAI) {
1261 void MyGCPrinter::finishAssembly(std::ostream &OS, AsmPrinter &AP,
1262 const TargetAsmInfo &TAI) {
1263 // Set up for emitting addresses.
1264 const char *AddressDirective;
1265 int AddressAlignLog;
1266 if (AP.TM.getTargetData()->getPointerSize() == sizeof(int32_t)) {
1267 AddressDirective = TAI.getData32bitsDirective();
1268 AddressAlignLog = 2;
1270 AddressDirective = TAI.getData64bitsDirective();
1271 AddressAlignLog = 3;
1274 // Put this in the data section.
1275 AP.SwitchToDataSection(TAI.getDataSection());
1277 // For each function...
1278 for (iterator FI = begin(), FE = end(); FI != FE; ++FI) {
1279 GCFunctionInfo &MD = **FI;
1281 // Emit this data structure:
1284 // int32_t PointCount;
1286 // void *SafePointAddress;
1287 // int32_t LiveCount;
1288 // int32_t LiveOffsets[LiveCount];
1289 // } Points[PointCount];
1290 // } __gcmap_<FUNCTIONNAME>;
1292 // Align to address width.
1293 AP.EmitAlignment(AddressAlignLog);
1295 // Emit the symbol by which the stack map entry can be found.
1297 Symbol += TAI.getGlobalPrefix();
1298 Symbol += "__gcmap_";
1299 Symbol += MD.getFunction().getName();
1300 if (const char *GlobalDirective = TAI.getGlobalDirective())
1301 OS << GlobalDirective << Symbol << "\n";
1302 OS << TAI.getGlobalPrefix() << Symbol << ":\n";
1305 AP.EmitInt32(MD.size());
1306 AP.EOL("safe point count");
1308 // And each safe point...
1309 for (GCFunctionInfo::iterator PI = MD.begin(),
1310 PE = MD.end(); PI != PE; ++PI) {
1311 // Align to address width.
1312 AP.EmitAlignment(AddressAlignLog);
1314 // Emit the address of the safe point.
1315 OS << AddressDirective
1316 << TAI.getPrivateGlobalPrefix() << "label" << PI->Num;
1317 AP.EOL("safe point address");
1319 // Emit the stack frame size.
1320 AP.EmitInt32(MD.getFrameSize());
1321 AP.EOL("stack frame size");
1323 // Emit the number of live roots in the function.
1324 AP.EmitInt32(MD.live_size(PI));
1325 AP.EOL("live root count");
1327 // And for each live root...
1328 for (GCFunctionInfo::live_iterator LI = MD.live_begin(PI),
1329 LE = MD.live_end(PI);
1331 // Print its offset within the stack frame.
1332 AP.EmitInt32(LI->StackOffset);
1333 AP.EOL("stack offset");
1344 <!-- *********************************************************************** -->
1346 <a name="references">References</a>
1348 <!-- *********************************************************************** -->
1352 <p><a name="appel89">[Appel89]</a> Runtime Tags Aren't Necessary. Andrew
1353 W. Appel. Lisp and Symbolic Computation 19(7):703-705, July 1989.</p>
1355 <p><a name="goldberg91">[Goldberg91]</a> Tag-free garbage collection for
1356 strongly typed programming languages. Benjamin Goldberg. ACM SIGPLAN
1359 <p><a name="tolmach94">[Tolmach94]</a> Tag-free garbage collection using
1360 explicit type parameters. Andrew Tolmach. Proceedings of the 1994 ACM
1361 conference on LISP and functional programming.</p>
1363 <p><a name="henderson02">[Henderson2002]</a> <a
1364 href="http://citeseer.ist.psu.edu/henderson02accurate.html">
1365 Accurate Garbage Collection in an Uncooperative Environment</a>.
1366 Fergus Henderson. International Symposium on Memory Management 2002.</p>
1371 <!-- *********************************************************************** -->
1375 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
1376 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
1377 <a href="http://validator.w3.org/check/referer"><img
1378 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
1380 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
1381 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
1382 Last modified: $Date$