1 =====================================
2 Garbage Collection with LLVM
3 =====================================
11 This document covers how to integrate LLVM into a compiler for a language which
12 supports garbage collection. **Note that LLVM itself does not provide a
13 garbage collector.** You must provide your own.
18 First, you should pick a collector strategy. LLVM includes a number of built
19 in ones, but you can also implement a loadable plugin with a custom definition.
20 Note that the collector strategy is a description of how LLVM should generate
21 code such that it interacts with your collector and runtime, not a description
22 of the collector itself.
24 Next, mark your generated functions as using your chosen collector strategy.
25 From c++, you can call:
29 F.setGC(<collector description name>);
32 This will produce IR like the following fragment:
36 define void @foo() gc "<collector description name>" { ... }
39 When generating LLVM IR for your functions, you will need to:
41 * Use ``@llvm.gcread`` and/or ``@llvm.gcwrite`` in place of standard load and
42 store instructions. These intrinsics are used to represent load and store
43 barriers. If you collector does not require such barriers, you can skip
46 * Use the memory allocation routines provided by your garbage collector's
49 * If your collector requires them, generate type maps according to your
50 runtime's binary interface. LLVM is not involved in the process. In
51 particular, the LLVM type system is not suitable for conveying such
52 information though the compiler.
54 * Insert any coordination code required for interacting with your collector.
55 Many collectors require running application code to periodically check a
56 flag and conditionally call a runtime function. This is often referred to
59 You will need to identify roots (i.e. references to heap objects your collector
60 needs to know about) in your generated IR, so that LLVM can encode them into
61 your final stack maps. Depending on the collector strategy chosen, this is
62 accomplished by using either the ''@llvm.gcroot'' intrinsics or an
63 ''gc.statepoint'' relocation sequence.
65 Don't forget to create a root for each intermediate value that is generated when
66 evaluating an expression. In ``h(f(), g())``, the result of ``f()`` could
67 easily be collected if evaluating ``g()`` triggers a collection.
69 Finally, you need to link your runtime library with the generated program
70 executable (for a static compiler) or ensure the appropriate symbols are
71 available for the runtime linker (for a JIT compiler).
77 What is Garbage Collection?
78 ---------------------------
80 Garbage collection is a widely used technique that frees the programmer from
81 having to know the lifetimes of heap objects, making software easier to produce
82 and maintain. Many programming languages rely on garbage collection for
83 automatic memory management. There are two primary forms of garbage collection:
84 conservative and accurate.
86 Conservative garbage collection often does not require any special support from
87 either the language or the compiler: it can handle non-type-safe programming
88 languages (such as C/C++) and does not require any special information from the
89 compiler. The `Boehm collector
90 <http://www.hpl.hp.com/personal/Hans_Boehm/gc/>`__ is an example of a
91 state-of-the-art conservative collector.
93 Accurate garbage collection requires the ability to identify all pointers in the
94 program at run-time (which requires that the source-language be type-safe in
95 most cases). Identifying pointers at run-time requires compiler support to
96 locate all places that hold live pointer variables at run-time, including the
97 :ref:`processor stack and registers <gcroot>`.
99 Conservative garbage collection is attractive because it does not require any
100 special compiler support, but it does have problems. In particular, because the
101 conservative garbage collector cannot *know* that a particular word in the
102 machine is a pointer, it cannot move live objects in the heap (preventing the
103 use of compacting and generational GC algorithms) and it can occasionally suffer
104 from memory leaks due to integer values that happen to point to objects in the
105 program. In addition, some aggressive compiler transformations can break
106 conservative garbage collectors (though these seem rare in practice).
108 Accurate garbage collectors do not suffer from any of these problems, but they
109 can suffer from degraded scalar optimization of the program. In particular,
110 because the runtime must be able to identify and update all pointers active in
111 the program, some optimizations are less effective. In practice, however, the
112 locality and performance benefits of using aggressive garbage collection
113 techniques dominates any low-level losses.
115 This document describes the mechanisms and interfaces provided by LLVM to
116 support accurate garbage collection.
121 LLVM's intermediate representation provides :ref:`garbage collection intrinsics
122 <gc_intrinsics>` that offer support for a broad class of collector models. For
123 instance, the intrinsics permit:
125 * semi-space collectors
127 * mark-sweep collectors
129 * generational collectors
131 * incremental collectors
133 * concurrent collectors
135 * cooperative collectors
139 We hope that the support built into the LLVM IR is sufficient to support a
140 broad class of garbage collected languages including Scheme, ML, Java, C#,
141 Perl, Python, Lua, Ruby, other scripting languages, and more.
143 Note that LLVM **does not itself provide a garbage collector** --- this should
144 be part of your language's runtime library. LLVM provides a framework for
145 describing the garbage collectors requirements to the compiler. In particular,
146 LLVM provides support for generating stack maps at call sites, polling for a
147 safepoint, and emitting load and store barriers. You can also extend LLVM -
148 possibly through a loadable :ref:`code generation plugins <plugin>` - to
149 generate code and data structures which conforms to the *binary interface*
150 specified by the *runtime library*. This is similar to the relationship between
151 LLVM and DWARF debugging info, for example. The difference primarily lies in
152 the lack of an established standard in the domain of garbage collection --- thus
153 the need for a flexible extension mechanism.
155 The aspects of the binary interface with which LLVM's GC support is
158 * Creation of GC safepoints within code where collection is allowed to execute
161 * Computation of the stack map. For each safe point in the code, object
162 references within the stack frame must be identified so that the collector may
163 traverse and perhaps update them.
165 * Write barriers when storing object references to the heap. These are commonly
166 used to optimize incremental scans in generational collectors.
168 * Emission of read barriers when loading object references. These are useful
169 for interoperating with concurrent collectors.
171 There are additional areas that LLVM does not directly address:
173 * Registration of global roots with the runtime.
175 * Registration of stack map entries with the runtime.
177 * The functions used by the program to allocate memory, trigger a collection,
180 * Computation or compilation of type maps, or registration of them with the
181 runtime. These are used to crawl the heap for object references.
183 In general, LLVM's support for GC does not include features which can be
184 adequately addressed with other features of the IR and does not specify a
185 particular binary interface. On the plus side, this means that you should be
186 able to integrate LLVM with an existing runtime. On the other hand, it can
187 have the effect of leaving a lot of work for the developer of a novel
188 language. We try to mitigate this by providing built in collector strategy
189 descriptions that can work with many common collector designs and easy
190 extension points. If you don't already have a specific binary interface
191 you need to support, we recommend trying to use one of these built in collector
199 This section describes the garbage collection facilities provided by the
200 :doc:`LLVM intermediate representation <LangRef>`. The exact behavior of these
201 IR features is specified by the selected :ref:`GC strategy description
204 Specifying GC code generation: ``gc "..."``
205 -------------------------------------------
209 define <returntype> @name(...) gc "name" { ... }
211 The ``gc`` function attribute is used to specify the desired GC strategy to the
212 compiler. Its programmatic equivalent is the ``setGC`` method of ``Function``.
214 Setting ``gc "name"`` on a function triggers a search for a matching subclass
215 of GCStrategy. Some collector strategies are built in. You can add others
216 using either the loadable plugin mechanism, or by patching your copy of LLVM.
217 It is the selected GC strategy which defines the exact nature of the code
218 generated to support GC. If none is found, the compiler will raise an error.
220 Specifying the GC style on a per-function basis allows LLVM to link together
221 programs that use different garbage collection algorithms (or none at all).
225 Identifying GC roots on the stack
226 ----------------------------------
228 LLVM currently supports two different mechanisms for describing references in
229 compiled code at safepoints. ``llvm.gcroot`` is the older mechanism;
230 ``gc.statepoint`` has been added more recently. At the moment, you can choose
231 either implementation (on a per :ref:`GC strategy <plugin>` basis). Longer
232 term, we will probably either migrate away from ``llvm.gcroot`` entirely, or
233 substantially merge their implementations. Note that most new development
234 work is focused on ``gc.statepoint``.
236 Using ``gc.statepoint``
237 ^^^^^^^^^^^^^^^^^^^^^^^^
238 :doc:`This page <Statepoints>` contains detailed documentation for
241 Using ``llvm.gcwrite``
242 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
246 void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
248 The ``llvm.gcroot`` intrinsic is used to inform LLVM that a stack variable
249 references an object on the heap and is to be tracked for garbage collection.
250 The exact impact on generated code is specified by a :ref:`compiler plugin
251 <plugin>`. All calls to ``llvm.gcroot`` **must** reside inside the first basic
254 The first argument **must** be a value referring to an alloca instruction or a
255 bitcast of an alloca. The second contains a pointer to metadata that should be
256 associated with the pointer, and **must** be a constant or global value
257 address. If your target collector uses tags, use a null pointer for metadata.
259 A compiler which performs manual SSA construction **must** ensure that SSA
260 values representing GC references are stored in to the alloca passed to the
261 respective ``gcroot`` before every call site and reloaded after every call.
262 A compiler which uses mem2reg to raise imperative code using ``alloca`` into
263 SSA form need only add a call to ``@llvm.gcroot`` for those variables which
264 are pointers into the GC heap.
266 It is also important to mark intermediate values with ``llvm.gcroot``. For
267 example, consider ``h(f(), g())``. Beware leaking the result of ``f()`` in the
268 case that ``g()`` triggers a collection. Note, that stack variables must be
269 initialized and marked with ``llvm.gcroot`` in function's prologue.
271 The ``%metadata`` argument can be used to avoid requiring heap objects to have
272 'isa' pointers or tag bits. [Appel89_, Goldberg91_, Tolmach94_] If specified,
273 its value will be tracked along with the location of the pointer in the stack
276 Consider the following fragment of Java code:
281 Object X; // A null-initialized reference to an object
285 This block (which may be located in the middle of a function or in a loop nest),
286 could be compiled to this LLVM code:
291 ;; In the entry block for the function, allocate the
292 ;; stack space for X, which is an LLVM pointer.
295 ;; Tell LLVM that the stack space is a stack root.
296 ;; Java has type-tags on objects, so we pass null as metadata.
297 %tmp = bitcast %Object** %X to i8**
298 call void @llvm.gcroot(i8** %tmp, i8* null)
301 ;; "CodeBlock" is the block corresponding to the start
302 ;; of the scope above.
304 ;; Java null-initializes pointers.
305 store %Object* null, %Object** %X
309 ;; As the pointer goes out of scope, store a null value into
310 ;; it, to indicate that the value is no longer live.
311 store %Object* null, %Object** %X
314 Reading and writing references in the heap
315 ------------------------------------------
317 Some collectors need to be informed when the mutator (the program that needs
318 garbage collection) either reads a pointer from or writes a pointer to a field
319 of a heap object. The code fragments inserted at these points are called *read
320 barriers* and *write barriers*, respectively. The amount of code that needs to
321 be executed is usually quite small and not on the critical path of any
322 computation, so the overall performance impact of the barrier is tolerable.
324 Barriers often require access to the *object pointer* rather than the *derived
325 pointer* (which is a pointer to the field within the object). Accordingly,
326 these intrinsics take both pointers as separate arguments for completeness. In
327 this snippet, ``%object`` is the object pointer, and ``%derived`` is the derived
333 %class.Array = type { %class.Object, i32, [0 x %class.Object*] }
336 ;; Load the object pointer from a gcroot.
337 %object = load %class.Array** %object_addr
339 ;; Compute the derived pointer.
340 %derived = getelementptr %object, i32 0, i32 2, i32 %n
342 LLVM does not enforce this relationship between the object and derived pointer
343 (although a particular :ref:`collector strategy <plugin>` might). However, it
344 would be an unusual collector that violated it.
346 The use of these intrinsics is naturally optional if the target GC does not
347 require the corresponding barrier. The GC strategy used with such a collector
348 should replace the intrinsic calls with the corresponding ``load`` or
349 ``store`` instruction if they are used.
351 One known deficiency with the current design is that the barrier intrinsics do
352 not include the size or alignment of the underlying operation performed. It is
353 currently assumed that the operation is of pointer size and the alignment is
354 assumed to be the target machine's default alignment.
356 Write barrier: ``llvm.gcwrite``
357 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
361 void @llvm.gcwrite(i8* %value, i8* %object, i8** %derived)
363 For write barriers, LLVM provides the ``llvm.gcwrite`` intrinsic function. It
364 has exactly the same semantics as a non-volatile ``store`` to the derived
365 pointer (the third argument). The exact code generated is specified by the
366 Function's selected :ref:`GC strategy <plugin>`.
368 Many important algorithms require write barriers, including generational and
369 concurrent collectors. Additionally, write barriers could be used to implement
372 Read barrier: ``llvm.gcread``
373 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
377 i8* @llvm.gcread(i8* %object, i8** %derived)
379 For read barriers, LLVM provides the ``llvm.gcread`` intrinsic function. It has
380 exactly the same semantics as a non-volatile ``load`` from the derived pointer
381 (the second argument). The exact code generated is specified by the Function's
382 selected :ref:`GC strategy <plugin>`.
384 Read barriers are needed by fewer algorithms than write barriers, and may have a
385 greater performance impact since pointer reads are more frequent than writes.
392 LLVM includes built in support for several varieties of garbage collectors.
395 ----------------------
397 To use this collector strategy, mark your functions with:
401 F.setGC("shadow-stack");
403 Unlike many GC algorithms which rely on a cooperative code generator to compile
404 stack maps, this algorithm carefully maintains a linked list of stack roots
405 [:ref:`Henderson2002 <henderson02>`]. This so-called "shadow stack" mirrors the
406 machine stack. Maintaining this data structure is slower than using a stack map
407 compiled into the executable as constant data, but has a significant portability
408 advantage because it requires no special support from the target code generator,
409 and does not require tricky platform-specific code to crawl the machine stack.
411 The tradeoff for this simplicity and portability is:
413 * High overhead per function call.
417 Still, it's an easy way to get started. After your compiler and runtime are up
418 and running, writing a :ref:`plugin <plugin>` will allow you to take advantage
419 of :ref:`more advanced GC features <collector-algos>` of LLVM in order to
423 The shadow stack doesn't imply a memory allocation algorithm. A semispace
424 collector or building atop ``malloc`` are great places to start, and can be
425 implemented with very little code.
427 When it comes time to collect, however, your runtime needs to traverse the stack
428 roots, and for this it needs to integrate with the shadow stack. Luckily, doing
429 so is very simple. (This code is heavily commented to help you understand the
430 data structure, but there are only 20 lines of meaningful code.)
434 /// @brief The map for a single function's stack frame. One of these is
435 /// compiled as constant data into the executable for each function.
437 /// Storage of metadata values is elided if the %metadata parameter to
438 /// @llvm.gcroot is null.
440 int32_t NumRoots; //< Number of roots in stack frame.
441 int32_t NumMeta; //< Number of metadata entries. May be < NumRoots.
442 const void *Meta[0]; //< Metadata for each root.
445 /// @brief A link in the dynamic shadow stack. One of these is embedded in
446 /// the stack frame of each function on the call stack.
448 StackEntry *Next; //< Link to next stack entry (the caller's).
449 const FrameMap *Map; //< Pointer to constant FrameMap.
450 void *Roots[0]; //< Stack roots (in-place array).
453 /// @brief The head of the singly-linked list of StackEntries. Functions push
454 /// and pop onto this in their prologue and epilogue.
456 /// Since there is only a global list, this technique is not threadsafe.
457 StackEntry *llvm_gc_root_chain;
459 /// @brief Calls Visitor(root, meta) for each GC root on the stack.
460 /// root and meta are exactly the values passed to
463 /// Visitor could be a function to recursively mark live objects. Or it
464 /// might copy them to another heap or generation.
466 /// @param Visitor A function to invoke for every GC root on the stack.
467 void visitGCRoots(void (*Visitor)(void **Root, const void *Meta)) {
468 for (StackEntry *R = llvm_gc_root_chain; R; R = R->Next) {
471 // For roots [0, NumMeta), the metadata pointer is in the FrameMap.
472 for (unsigned e = R->Map->NumMeta; i != e; ++i)
473 Visitor(&R->Roots[i], R->Map->Meta[i]);
475 // For roots [NumMeta, NumRoots), the metadata pointer is null.
476 for (unsigned e = R->Map->NumRoots; i != e; ++i)
477 Visitor(&R->Roots[i], NULL);
482 The 'Erlang' and 'Ocaml' GCs
483 -----------------------------
485 LLVM ships with two example collectors which leverage the ''gcroot''
486 mechanisms. To our knowledge, these are not actually used by any language
487 runtime, but they do provide a reasonable starting point for someone interested
488 in writing an ''gcroot' compatible GC plugin. In particular, these are the
489 only in tree examples of how to produce a custom binary stack map format using
490 a ''gcroot'' strategy.
492 As there names imply, the binary format produced is intended to model that
493 used by the Erlang and OCaml compilers respectively.
496 The Statepoint Example GC
497 -------------------------
501 F.setGC("statepoint-example");
503 This GC provides an example of how one might use the infrastructure provided
504 by ''gc.statepoint''.
510 If none of the built in GC strategy descriptions met your needs above, you will
511 need to define a custom GCStrategy and possibly, a custom LLVM pass to perform
512 lowering. Your best example of where to start defining a custom GCStrategy
513 would be to look at one of the built in strategies.
515 You may be able to structure this additional code as a loadable plugin library.
516 Loadable plugins are sufficient if all you need is to enable a different
517 combination of built in functionality, but if you need to provide a custom
518 lowering pass, you will need to build a patched version of LLVM. If you think
519 you need a patched build, please ask for advice on llvm-dev. There may be an
520 easy way we can extend the support to make it work for your use case without
521 requiring a custom build.
523 Collector Requirements
524 ----------------------
526 You should be able to leverage any existing collector library that includes the following elements:
528 #. A memory allocator which exposes an allocation function your compiled
531 #. A binary format for the stack map. A stack map describes the location
532 of references at a safepoint and is used by precise collectors to identify
533 references within a stack frame on the machine stack. Note that collectors
534 which conservatively scan the stack don't require such a structure.
536 #. A stack crawler to discover functions on the call stack, and enumerate the
537 references listed in the stack map for each call site.
539 #. A mechanism for identifying references in global locations (e.g. global
542 #. If you collector requires them, an LLVM IR implementation of your collectors
543 load and store barriers. Note that since many collectors don't require
544 barriers at all, LLVM defaults to lowering such barriers to normal loads
545 and stores unless you arrange otherwise.
548 Implementing a collector plugin
549 -------------------------------
551 User code specifies which GC code generation to use with the ``gc`` function
552 attribute or, equivalently, with the ``setGC`` method of ``Function``.
554 To implement a GC plugin, it is necessary to subclass ``llvm::GCStrategy``,
555 which can be accomplished in a few lines of boilerplate code. LLVM's
556 infrastructure provides access to several important algorithms. For an
557 uncontroversial collector, all that remains may be to compile LLVM's computed
558 stack map to assembly code (using the binary representation expected by the
559 runtime library). This can be accomplished in about 100 lines of code.
561 This is not the appropriate place to implement a garbage collected heap or a
562 garbage collector itself. That code should exist in the language's runtime
563 library. The compiler plugin is responsible for generating code which conforms
564 to the binary interface defined by library, most essentially the :ref:`stack map
567 To subclass ``llvm::GCStrategy`` and register it with the compiler:
571 // lib/MyGC/MyGC.cpp - Example LLVM GC plugin
573 #include "llvm/CodeGen/GCStrategy.h"
574 #include "llvm/CodeGen/GCMetadata.h"
575 #include "llvm/Support/Compiler.h"
577 using namespace llvm;
580 class LLVM_LIBRARY_VISIBILITY MyGC : public GCStrategy {
585 GCRegistry::Add<MyGC>
586 X("mygc", "My bespoke garbage collector.");
589 This boilerplate collector does nothing. More specifically:
591 * ``llvm.gcread`` calls are replaced with the corresponding ``load``
594 * ``llvm.gcwrite`` calls are replaced with the corresponding ``store``
597 * No safe points are added to the code.
599 * The stack map is not compiled into the executable.
601 Using the LLVM makefiles, this code
602 can be compiled as a plugin using a simple makefile:
612 include $(LEVEL)/Makefile.common
614 Once the plugin is compiled, code using it may be compiled using ``llc
615 -load=MyGC.so`` (though MyGC.so may have some other platform-specific
621 define void @f() gc "mygc" {
625 $ llvm-as < sample.ll | llc -load=MyGC.so
627 It is also possible to statically link the collector plugin into tools, such as
628 a language-specific compiler front-end.
632 Overview of available features
633 ------------------------------
635 ``GCStrategy`` provides a range of features through which a plugin may do useful
636 work. Some of these are callbacks, some are algorithms that can be enabled,
637 disabled, or customized. This matrix summarizes the supported (and planned)
638 features and correlates them with the collection techniques which typically
641 .. |v| unicode:: 0x2714
644 .. |x| unicode:: 0x2718
647 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
648 | Algorithm | Done | Shadow | refcount | mark- | copying | incremental | threaded | concurrent |
649 | | | stack | | sweep | | | | |
650 +============+======+========+==========+=======+=========+=============+==========+============+
651 | stack map | |v| | | | |x| | |x| | |x| | |x| | |x| |
652 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
653 | initialize | |v| | |x| | |x| | |x| | |x| | |x| | |x| | |x| |
654 | roots | | | | | | | | |
655 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
656 | derived | NO | | | | | | **N**\* | **N**\* |
657 | pointers | | | | | | | | |
658 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
659 | **custom | |v| | | | | | | | |
660 | lowering** | | | | | | | | |
661 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
662 | *gcroot* | |v| | |x| | |x| | | | | | |
663 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
664 | *gcwrite* | |v| | | |x| | | | |x| | | |x| |
665 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
666 | *gcread* | |v| | | | | | | | |x| |
667 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
668 | **safe | | | | | | | | |
669 | points** | | | | | | | | |
670 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
671 | *in | |v| | | | |x| | |x| | |x| | |x| | |x| |
672 | calls* | | | | | | | | |
673 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
674 | *before | |v| | | | | | | |x| | |x| |
675 | calls* | | | | | | | | |
676 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
677 | *for | NO | | | | | | **N** | **N** |
678 | loops* | | | | | | | | |
679 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
680 | *before | |v| | | | | | | |x| | |x| |
681 | escape* | | | | | | | | |
682 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
683 | emit code | NO | | | | | | **N** | **N** |
684 | at safe | | | | | | | | |
685 | points | | | | | | | | |
686 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
687 | **output** | | | | | | | | |
688 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
689 | *assembly* | |v| | | | |x| | |x| | |x| | |x| | |x| |
690 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
691 | *JIT* | NO | | | **?** | **?** | **?** | **?** | **?** |
692 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
693 | *obj* | NO | | | **?** | **?** | **?** | **?** | **?** |
694 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
695 | live | NO | | | **?** | **?** | **?** | **?** | **?** |
696 | analysis | | | | | | | | |
697 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
698 | register | NO | | | **?** | **?** | **?** | **?** | **?** |
699 | map | | | | | | | | |
700 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
701 | \* Derived pointers only pose a hasard to copying collections. |
702 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
703 | **?** denotes a feature which could be utilized if available. |
704 +------------+------+--------+----------+-------+---------+-------------+----------+------------+
706 To be clear, the collection techniques above are defined as:
709 The mutator carefully maintains a linked list of stack roots.
712 The mutator maintains a reference count for each object and frees an object
713 when its count falls to zero.
716 When the heap is exhausted, the collector marks reachable objects starting
717 from the roots, then deallocates unreachable objects in a sweep phase.
720 As reachability analysis proceeds, the collector copies objects from one heap
721 area to another, compacting them in the process. Copying collectors enable
722 highly efficient "bump pointer" allocation and can improve locality of
726 (Including generational collectors.) Incremental collectors generally have all
727 the properties of a copying collector (regardless of whether the mature heap
728 is compacting), but bring the added complexity of requiring write barriers.
731 Denotes a multithreaded mutator; the collector must still stop the mutator
732 ("stop the world") before beginning reachability analysis. Stopping a
733 multithreaded mutator is a complicated problem. It generally requires highly
734 platform-specific code in the runtime, and the production of carefully
735 designed machine code at safe points.
738 In this technique, the mutator and the collector run concurrently, with the
739 goal of eliminating pause times. In a *cooperative* collector, the mutator
740 further aids with collection should a pause occur, allowing collection to take
741 advantage of multiprocessor hosts. The "stop the world" problem of threaded
742 collectors is generally still present to a limited extent. Sophisticated
743 marking algorithms are necessary. Read barriers may be necessary.
745 As the matrix indicates, LLVM's garbage collection infrastructure is already
746 suitable for a wide variety of collectors, but does not currently extend to
747 multithreaded programs. This will be added in the future as there is
755 LLVM automatically computes a stack map. One of the most important features
756 of a ``GCStrategy`` is to compile this information into the executable in
757 the binary representation expected by the runtime library.
759 The stack map consists of the location and identity of each GC root in the
760 each function in the module. For each root:
762 * ``RootNum``: The index of the root.
764 * ``StackOffset``: The offset of the object relative to the frame pointer.
766 * ``RootMetadata``: The value passed as the ``%metadata`` parameter to the
767 ``@llvm.gcroot`` intrinsic.
769 Also, for the function as a whole:
771 * ``getFrameSize()``: The overall size of the function's initial stack frame,
772 not accounting for any dynamic allocation.
774 * ``roots_size()``: The count of roots in the function.
776 To access the stack map, use ``GCFunctionMetadata::roots_begin()`` and
777 -``end()`` from the :ref:`GCMetadataPrinter <assembly>`:
781 for (iterator I = begin(), E = end(); I != E; ++I) {
782 GCFunctionInfo *FI = *I;
783 unsigned FrameSize = FI->getFrameSize();
784 size_t RootCount = FI->roots_size();
786 for (GCFunctionInfo::roots_iterator RI = FI->roots_begin(),
787 RE = FI->roots_end();
789 int RootNum = RI->Num;
790 int RootStackOffset = RI->StackOffset;
791 Constant *RootMetadata = RI->Metadata;
795 If the ``llvm.gcroot`` intrinsic is eliminated before code generation by a
796 custom lowering pass, LLVM will compute an empty stack map. This may be useful
797 for collector plugins which implement reference counting or a shadow stack.
801 Initializing roots to null: ``InitRoots``
802 -----------------------------------------
810 When set, LLVM will automatically initialize each root to ``null`` upon entry to
811 the function. This prevents the GC's sweep phase from visiting uninitialized
812 pointers, which will almost certainly cause it to crash. This initialization
813 occurs before custom lowering, so the two may be used together.
815 Since LLVM does not yet compute liveness information, there is no means of
816 distinguishing an uninitialized stack root from an initialized one. Therefore,
817 this feature should be used by all GC plugins. It is enabled by default.
819 Custom lowering of intrinsics: ``CustomRoots``, ``CustomReadBarriers``, and ``CustomWriteBarriers``
820 ---------------------------------------------------------------------------------------------------
822 For GCs which use barriers or unusual treatment of stack roots, these
823 flags allow the collector to perform arbitrary transformations of the
828 class MyGC : public GCStrategy {
832 CustomReadBarriers = true;
833 CustomWriteBarriers = true;
837 If any of these flags are set, LLVM suppresses its default lowering for
838 the corresponding intrinsics. Instead, you must provide a custom Pass
839 which lowers the intrinsics as desired. If you have opted in to custom
840 lowering of a particular intrinsic your pass **must** eliminate all
841 instances of the corresponding intrinsic in functions which opt in to
842 your GC. The best example of such a pass is the ShadowStackGC and it's
843 ShadowStackGCLowering pass.
845 There is currently no way to register such a custom lowering pass
846 without building a custom copy of LLVM.
850 Generating safe points: ``NeededSafePoints``
851 --------------------------------------------
853 LLVM can compute four kinds of safe points:
858 /// PointKind - The type of a collector-safe point.
861 Loop, //< Instr is a loop (backwards branch).
862 Return, //< Instr is a return instruction.
863 PreCall, //< Instr is a call instruction.
864 PostCall //< Instr is the return address of a call.
868 A collector can request any combination of the four by setting the
869 ``NeededSafePoints`` mask:
874 NeededSafePoints = 1 << GC::Loop
880 It can then use the following routines to access safe points.
884 for (iterator I = begin(), E = end(); I != E; ++I) {
885 GCFunctionInfo *MD = *I;
886 size_t PointCount = MD->size();
888 for (GCFunctionInfo::iterator PI = MD->begin(),
889 PE = MD->end(); PI != PE; ++PI) {
890 GC::PointKind PointKind = PI->Kind;
891 unsigned PointNum = PI->Num;
895 Almost every collector requires ``PostCall`` safe points, since these correspond
896 to the moments when the function is suspended during a call to a subroutine.
898 Threaded programs generally require ``Loop`` safe points to guarantee that the
899 application will reach a safe point within a bounded amount of time, even if it
900 is executing a long-running loop which contains no function calls.
902 Threaded collectors may also require ``Return`` and ``PreCall`` safe points to
903 implement "stop the world" techniques using self-modifying code, where it is
904 important that the program not exit the function without reaching a safe point
905 (because only the topmost function has been patched).
909 Emitting assembly code: ``GCMetadataPrinter``
910 ---------------------------------------------
912 LLVM allows a plugin to print arbitrary assembly code before and after the rest
913 of a module's assembly code. At the end of the module, the GC can compile the
914 LLVM stack map into assembly code. (At the beginning, this information is not
917 Since AsmWriter and CodeGen are separate components of LLVM, a separate abstract
918 base class and registry is provided for printing assembly code, the
919 ``GCMetadaPrinter`` and ``GCMetadataPrinterRegistry``. The AsmWriter will look
920 for such a subclass if the ``GCStrategy`` sets ``UsesMetadata``:
928 This separation allows JIT-only clients to be smaller.
930 Note that LLVM does not currently have analogous APIs to support code generation
931 in the JIT, nor using the object writers.
935 // lib/MyGC/MyGCPrinter.cpp - Example LLVM GC printer
937 #include "llvm/CodeGen/GCMetadataPrinter.h"
938 #include "llvm/Support/Compiler.h"
940 using namespace llvm;
943 class LLVM_LIBRARY_VISIBILITY MyGCPrinter : public GCMetadataPrinter {
945 virtual void beginAssembly(AsmPrinter &AP);
947 virtual void finishAssembly(AsmPrinter &AP);
950 GCMetadataPrinterRegistry::Add<MyGCPrinter>
951 X("mygc", "My bespoke garbage collector.");
954 The collector should use ``AsmPrinter`` to print portable assembly code. The
955 collector itself contains the stack map for the entire module, and may access
956 the ``GCFunctionInfo`` using its own ``begin()`` and ``end()`` methods. Here's
961 #include "llvm/CodeGen/AsmPrinter.h"
962 #include "llvm/IR/Function.h"
963 #include "llvm/IR/DataLayout.h"
964 #include "llvm/Target/TargetAsmInfo.h"
965 #include "llvm/Target/TargetMachine.h"
967 void MyGCPrinter::beginAssembly(AsmPrinter &AP) {
971 void MyGCPrinter::finishAssembly(AsmPrinter &AP) {
972 MCStreamer &OS = AP.OutStreamer;
973 unsigned IntPtrSize = AP.TM.getSubtargetImpl()->getDataLayout()->getPointerSize();
975 // Put this in the data section.
976 OS.SwitchSection(AP.getObjFileLowering().getDataSection());
978 // For each function...
979 for (iterator FI = begin(), FE = end(); FI != FE; ++FI) {
980 GCFunctionInfo &MD = **FI;
982 // A compact GC layout. Emit this data structure:
985 // int32_t PointCount;
986 // void *SafePointAddress[PointCount];
987 // int32_t StackFrameSize; // in words
988 // int32_t StackArity;
989 // int32_t LiveCount;
990 // int32_t LiveOffsets[LiveCount];
991 // } __gcmap_<FUNCTIONNAME>;
993 // Align to address width.
994 AP.EmitAlignment(IntPtrSize == 4 ? 2 : 3);
997 OS.AddComment("safe point count");
998 AP.EmitInt32(MD.size());
1000 // And each safe point...
1001 for (GCFunctionInfo::iterator PI = MD.begin(),
1002 PE = MD.end(); PI != PE; ++PI) {
1003 // Emit the address of the safe point.
1004 OS.AddComment("safe point address");
1005 MCSymbol *Label = PI->Label;
1006 AP.EmitLabelPlusOffset(Label/*Hi*/, 0/*Offset*/, 4/*Size*/);
1009 // Stack information never change in safe points! Only print info from the
1011 GCFunctionInfo::iterator PI = MD.begin();
1013 // Emit the stack frame size.
1014 OS.AddComment("stack frame size (in words)");
1015 AP.EmitInt32(MD.getFrameSize() / IntPtrSize);
1017 // Emit stack arity, i.e. the number of stacked arguments.
1018 unsigned RegisteredArgs = IntPtrSize == 4 ? 5 : 6;
1019 unsigned StackArity = MD.getFunction().arg_size() > RegisteredArgs ?
1020 MD.getFunction().arg_size() - RegisteredArgs : 0;
1021 OS.AddComment("stack arity");
1022 AP.EmitInt32(StackArity);
1024 // Emit the number of live roots in the function.
1025 OS.AddComment("live root count");
1026 AP.EmitInt32(MD.live_size(PI));
1028 // And for each live root...
1029 for (GCFunctionInfo::live_iterator LI = MD.live_begin(PI),
1030 LE = MD.live_end(PI);
1032 // Emit live root's offset within the stack frame.
1033 OS.AddComment("stack index (offset / wordsize)");
1034 AP.EmitInt32(LI->StackOffset);
1044 [Appel89] Runtime Tags Aren't Necessary. Andrew W. Appel. Lisp and Symbolic
1045 Computation 19(7):703-705, July 1989.
1049 [Goldberg91] Tag-free garbage collection for strongly typed programming
1050 languages. Benjamin Goldberg. ACM SIGPLAN PLDI'91.
1054 [Tolmach94] Tag-free garbage collection using explicit type parameters. Andrew
1055 Tolmach. Proceedings of the 1994 ACM conference on LISP and functional
1060 [Henderson2002] `Accurate Garbage Collection in an Uncooperative Environment
1061 <http://citeseer.ist.psu.edu/henderson02accurate.html>`__