1 =====================================
2 Garbage Collection Safepoints in LLVM
3 =====================================
12 This document describes a set of experimental extensions to LLVM. Use
13 with caution. Because the intrinsics have experimental status,
14 compatibility across LLVM releases is not guaranteed.
16 LLVM currently supports an alternate mechanism for conservative
17 garbage collection support using the ``gcroot`` intrinsic. The mechanism
18 described here shares little in common with the alternate ``gcroot``
19 implementation and it is hoped that this mechanism will eventually
20 replace the gc_root mechanism.
25 To collect dead objects, garbage collectors must be able to identify
26 any references to objects contained within executing code, and,
27 depending on the collector, potentially update them. The collector
28 does not need this information at all points in code - that would make
29 the problem much harder - but only at well-defined points in the
30 execution known as 'safepoints' For most collectors, it is sufficient
31 to track at least one copy of each unique pointer value. However, for
32 a collector which wishes to relocate objects directly reachable from
33 running code, a higher standard is required.
35 One additional challenge is that the compiler may compute intermediate
36 results ("derived pointers") which point outside of the allocation or
37 even into the middle of another allocation. The eventual use of this
38 intermediate value must yield an address within the bounds of the
39 allocation, but such "exterior derived pointers" may be visible to the
40 collector. Given this, a garbage collector can not safely rely on the
41 runtime value of an address to indicate the object it is associated
42 with. If the garbage collector wishes to move any object, the
43 compiler must provide a mapping, for each pointer, to an indication of
46 To simplify the interaction between a collector and the compiled code,
47 most garbage collectors are organized in terms of three abstractions:
48 load barriers, store barriers, and safepoints.
50 #. A load barrier is a bit of code executed immediately after the
51 machine load instruction, but before any use of the value loaded.
52 Depending on the collector, such a barrier may be needed for all
53 loads, merely loads of a particular type (in the original source
54 language), or none at all.
56 #. Analogously, a store barrier is a code fragement that runs
57 immediately before the machine store instruction, but after the
58 computation of the value stored. The most common use of a store
59 barrier is to update a 'card table' in a generational garbage
62 #. A safepoint is a location at which pointers visible to the compiled
63 code (i.e. currently in registers or on the stack) are allowed to
64 change. After the safepoint completes, the actual pointer value
65 may differ, but the 'object' (as seen by the source language)
68 Note that the term 'safepoint' is somewhat overloaded. It refers to
69 both the location at which the machine state is parsable and the
70 coordination protocol involved in bring application threads to a
71 point at which the collector can safely use that information. The
72 term "statepoint" as used in this document refers exclusively to the
75 This document focuses on the last item - compiler support for
76 safepoints in generated code. We will assume that an outside
77 mechanism has decided where to place safepoints. From our
78 perspective, all safepoints will be function calls. To support
79 relocation of objects directly reachable from values in compiled code,
80 the collector must be able to:
82 #. identify every copy of a pointer (including copies introduced by
83 the compiler itself) at the safepoint,
84 #. identify which object each pointer relates to, and
85 #. potentially update each of those copies.
87 This document describes the mechanism by which an LLVM based compiler
88 can provide this information to a language runtime/collector, and
89 ensure that all pointers can be read and updated if desired. The
90 heart of the approach is to construct (or rewrite) the IR in a manner
91 where the possible updates performed by the garbage collector are
92 explicitly visible in the IR. Doing so requires that we:
94 #. create a new SSA value for each potentially relocated pointer, and
95 ensure that no uses of the original (non relocated) value is
96 reachable after the safepoint,
97 #. specify the relocation in a way which is opaque to the compiler to
98 ensure that the optimizer can not introduce new uses of an
99 unrelocated value after a statepoint. This prevents the optimizer
100 from performing unsound optimizations.
101 #. recording a mapping of live pointers (and the allocation they're
102 associated with) for each statepoint.
104 At the most abstract level, inserting a safepoint can be thought of as
105 replacing a call instruction with a call to a multiple return value
106 function which both calls the original target of the call, returns
107 it's result, and returns updated values for any live pointers to
108 garbage collected objects.
110 Note that the task of identifying all live pointers to garbage
111 collected values, transforming the IR to expose a pointer giving the
112 base object for every such live pointer, and inserting all the
113 intrinsics correctly is explicitly out of scope for this document.
114 The recommended approach is to use the :ref:`utility passes
115 <statepoint-utilities>` described below.
117 This abstract function call is concretely represented by a sequence of
118 intrinsic calls known collectively as a "statepoint relocation sequence".
120 Let's consider a simple call in LLVM IR:
124 define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj)
125 gc "statepoint-example" {
127 ret i8 addrspace(1)* %obj
130 Depending on our language we may need to allow a safepoint during the execution
131 of ``foo``. If so, we need to let the collector update local values in the
132 current frame. If we don't, we'll be accessing a potential invalid reference
133 once we eventually return from the call.
135 In this example, we need to relocate the SSA value ``%obj``. Since we can't
136 actually change the value in the SSA value ``%obj``, we need to introduce a new
137 SSA value ``%obj.relocated`` which represents the potentially changed value of
138 ``%obj`` after the safepoint and update any following uses appropriately. The
139 resulting relocation sequence is:
143 define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj)
144 gc "statepoint-example" {
145 %0 = call i32 (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 0, i8 addrspace(1)* %obj)
146 %obj.relocated = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(i32 %0, i32 7, i32 7)
147 ret i8 addrspace(1)* %obj.relocated
150 Ideally, this sequence would have been represented as a M argument, N
151 return value function (where M is the number of values being
152 relocated + the original call arguments and N is the original return
153 value + each relocated value), but LLVM does not easily support such a
156 Instead, the statepoint intrinsic marks the actual site of the
157 safepoint or statepoint. The statepoint returns a token value (which
158 exists only at compile time). To get back the original return value
159 of the call, we use the ``gc.result`` intrinsic. To get the relocation
160 of each pointer in turn, we use the ``gc.relocate`` intrinsic with the
161 appropriate index. Note that both the ``gc.relocate`` and ``gc.result`` are
162 tied to the statepoint. The combination forms a "statepoint relocation
163 sequence" and represents the entitety of a parseable call or 'statepoint'.
165 When lowered, this example would generate the following x86 assembly:
174 movq (%rsp), %rax # This load is redundant (oops!)
178 Each of the potentially relocated values has been spilled to the
179 stack, and a record of that location has been recorded to the
180 :ref:`Stack Map section <stackmap-section>`. If the garbage collector
181 needs to update any of these pointers during the call, it knows
182 exactly what to change.
184 The relevant parts of the StackMap section for our example are:
188 # This describes the call site
189 # Stack Maps: callsite 2882400000
193 # .. 8 entries skipped ..
194 # This entry describes the spill slot which is directly addressable
195 # off RSP with offset 0. Given the value was spilled with a pushq,
197 # Stack Maps: Loc 8: Direct RSP [encoding: .byte 2, .byte 8, .short 7, .int 0]
203 This example was taken from the tests for the :ref:`RewriteStatepointsForGC` utility pass. As such, it's full StackMap can be easily examined with the following command.
207 opt -rewrite-statepoints-for-gc test/Transforms/RewriteStatepointsForGC/basics.ll -S | llc -debug-only=stackmaps
209 Base & Derived Pointers
210 ^^^^^^^^^^^^^^^^^^^^^^^
212 A "base pointer" is one which points to the starting address of an allocation
213 (object). A "derived pointer" is one which is offset from a base pointer by
214 some amount. When relocating objects, a garbage collector needs to be able
215 to relocate each derived pointer associated with an allocation to the same
216 offset from the new address.
218 "Interior derived pointers" remain within the bounds of the allocation
219 they're associated with. As a result, the base object can be found at
220 runtime provided the bounds of allocations are known to the runtime system.
222 "Exterior derived pointers" are outside the bounds of the associated object;
223 they may even fall within *another* allocations address range. As a result,
224 there is no way for a garbage collector to determine which allocation they
225 are associated with at runtime and compiler support is needed.
227 The ``gc.relocate`` intrinsic supports an explicit operand for describing the
228 allocation associated with a derived pointer. This operand is frequently
229 referred to as the base operand, but does not strictly speaking have to be
230 a base pointer, but it does need to lie within the bounds of the associated
231 allocation. Some collectors may require that the operand be an actual base
232 pointer rather than merely an internal derived pointer. Note that during
233 lowering both the base and derived pointer operands are required to be live
234 over the associated call safepoint even if the base is otherwise unused
237 If we extend our previous example to include a pointless derived pointer,
242 define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj)
243 gc "statepoint-example" {
244 %gep = getelementptr i8, i8 addrspace(1)* %obj, i64 20000
245 %token = call i32 (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 0, i8 addrspace(1)* %obj, i8 addrspace(1)* %gep)
246 %obj.relocated = call i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(i32 %token, i32 7, i32 7)
247 %gep.relocated = call i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(i32 %token, i32 7, i32 8)
248 %p = getelementptr i8, i8 addrspace(1)* %gep, i64 -20000
249 ret i8 addrspace(1)* %p
252 Note that in this example %p and %obj.relocate are the same address and we
253 could replace one with the other, potentially removing the derived pointer
254 from the live set at the safepoint entirely.
259 As a practical consideration, many garbage-collected systems allow code that is
260 collector-aware ("managed code") to call code that is not collector-aware
261 ("unmanaged code"). It is common that such calls must also be safepoints, since
262 it is desirable to allow the collector to run during the execution of
263 unmanaged code. Futhermore, it is common that coordinating the transition from
264 managed to unmanaged code requires extra code generation at the call site to
265 inform the collector of the transition. In order to support these needs, a
266 statepoint may be marked as a GC transition, and data that is necessary to
267 perform the transition (if any) may be provided as additional arguments to the
270 Note that although in many cases statepoints may be inferred to be GC
271 transitions based on the function symbols involved (e.g. a call from a
272 function with GC strategy "foo" to a function with GC strategy "bar"),
273 indirect calls that are also GC transitions must also be supported. This
274 requirement is the driving force behing the decision to require that GC
275 transitions are explicitly marked.
277 Let's revisit the sample given above, this time treating the call to ``@foo``
278 as a GC transition. Depending on our target, the transition code may need to
279 access some extra state in order to inform the collector of the transition.
280 Let's assume a hypothetical GC--somewhat unimaginatively named "hypothetical-gc"
281 --that requires that a TLS variable must be written to before and after a call
282 to unmanaged code. The resulting relocation sequence is:
286 @flag = thread_local global i32 0, align 4
288 define i8 addrspace(1)* @test1(i8 addrspace(1) *%obj)
289 gc "hypothetical-gc" {
291 %0 = call i32 (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* @foo, i32 0, i32 1, i32* @Flag, i32 0, i8 addrspace(1)* %obj)
292 %obj.relocated = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(i32 %0, i32 7, i32 7)
293 ret i8 addrspace(1)* %obj.relocated
296 During lowering, this will result in a instruction selection DAG that looks
303 GC_TRANSITION_START (lowered i32 *@Flag), SRCVALUE i32* Flag
305 GC_TRANSITION_END (lowered i32 *@Flag), SRCVALUE i32 *Flag
309 In order to generate the necessary transition code, the backend for each target
310 supported by "hypothetical-gc" must be modified to lower ``GC_TRANSITION_START``
311 and ``GC_TRANSITION_END`` nodes appropriately when the "hypothetical-gc"
312 strategy is in use for a particular function. Assuming that such lowering has
313 been added for X86, the generated assembly would be:
320 movl $1, %fs:Flag@TPOFF
322 movl $0, %fs:Flag@TPOFF
324 movq (%rsp), %rax # This load is redundant (oops!)
328 Note that the design as presented above is not fully implemented: in particular,
329 strategy-specific lowering is not present, and all GC transitions are emitted as
330 as single no-op before and after the call instruction. These no-ops are often
331 removed by the backend during dead machine instruction elimination.
337 'llvm.experimental.gc.statepoint' Intrinsic
338 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
346 @llvm.experimental.gc.statepoint(i64 <id>, i32 <num patch bytes>,
348 i64 <#call args>, i64 <flags>,
349 ... (call parameters),
350 i64 <# transition args>, ... (transition parameters),
351 i64 <# deopt args>, ... (deopt parameters),
357 The statepoint intrinsic represents a call which is parse-able by the
363 The 'id' operand is a constant integer that is reported as the ID
364 field in the generated stackmap. LLVM does not interpret this
365 parameter in any way and its meaning is up to the statepoint user to
366 decide. Note that LLVM is free to duplicate code containing
367 statepoint calls, and this may transform IR that had a unique 'id' per
368 lexical call to statepoint to IR that does not.
370 If 'num patch bytes' is non-zero then the call instruction
371 corresponding to the statepoint is not emitted and LLVM emits 'num
372 patch bytes' bytes of nops in its place. LLVM will emit code to
373 prepare the function arguments and retrieve the function return value
374 in accordance to the calling convention; the former before the nop
375 sequence and the latter after the nop sequence. It is expected that
376 the user will patch over the 'num patch bytes' bytes of nops with a
377 calling sequence specific to their runtime before executing the
378 generated machine code. There are no guarantees with respect to the
379 alignment of the nop sequence. Unlike :doc:`StackMaps` statepoints do
380 not have a concept of shadow bytes. Note that semantically the
381 statepoint still represents a call or invoke to 'target', and the nop
382 sequence after patching is expected to represent an operation
383 equivalent to a call or invoke to 'target'.
385 The 'target' operand is the function actually being called. The
386 target can be specified as either a symbolic LLVM function, or as an
387 arbitrary Value of appropriate function type. Note that the function
388 type must match the signature of the callee and the types of the 'call
389 parameters' arguments.
391 The '#call args' operand is the number of arguments to the actual
392 call. It must exactly match the number of arguments passed in the
393 'call parameters' variable length section.
395 The 'flags' operand is used to specify extra information about the
396 statepoint. This is currently only used to mark certain statepoints
397 as GC transitions. This operand is a 64-bit integer with the following
398 layout, where bit 0 is the least significant bit:
400 +-------+---------------------------------------------------+
402 +=======+===================================================+
403 | 0 | Set if the statepoint is a GC transition, cleared |
405 +-------+---------------------------------------------------+
406 | 1-63 | Reserved for future use; must be cleared. |
407 +-------+---------------------------------------------------+
409 The 'call parameters' arguments are simply the arguments which need to
410 be passed to the call target. They will be lowered according to the
411 specified calling convention and otherwise handled like a normal call
412 instruction. The number of arguments must exactly match what is
413 specified in '# call args'. The types must match the signature of
416 The 'transition parameters' arguments contain an arbitrary list of
417 Values which need to be passed to GC transition code. They will be
418 lowered and passed as operands to the appropriate GC_TRANSITION nodes
419 in the selection DAG. It is assumed that these arguments must be
420 available before and after (but not necessarily during) the execution
421 of the callee. The '# transition args' field indicates how many operands
422 are to be interpreted as 'transition parameters'.
424 The 'deopt parameters' arguments contain an arbitrary list of Values
425 which is meaningful to the runtime. The runtime may read any of these
426 values, but is assumed not to modify them. If the garbage collector
427 might need to modify one of these values, it must also be listed in
428 the 'gc pointer' argument list. The '# deopt args' field indicates
429 how many operands are to be interpreted as 'deopt parameters'.
431 The 'gc parameters' arguments contain every pointer to a garbage
432 collector object which potentially needs to be updated by the garbage
433 collector. Note that the argument list must explicitly contain a base
434 pointer for every derived pointer listed. The order of arguments is
435 unimportant. Unlike the other variable length parameter sets, this
436 list is not length prefixed.
441 A statepoint is assumed to read and write all memory. As a result,
442 memory operations can not be reordered past a statepoint. It is
443 illegal to mark a statepoint as being either 'readonly' or 'readnone'.
445 Note that legal IR can not perform any memory operation on a 'gc
446 pointer' argument of the statepoint in a location statically reachable
447 from the statepoint. Instead, the explicitly relocated value (from a
448 ``gc.relocate``) must be used.
450 'llvm.experimental.gc.result' Intrinsic
451 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
459 @llvm.experimental.gc.result(i32 %statepoint_token)
464 ``gc.result`` extracts the result of the original call instruction
465 which was replaced by the ``gc.statepoint``. The ``gc.result``
466 intrinsic is actually a family of three intrinsics due to an
467 implementation limitation. Other than the type of the return value,
468 the semantics are the same.
473 The first and only argument is the ``gc.statepoint`` which starts
474 the safepoint sequence of which this ``gc.result`` is a part.
475 Despite the typing of this as a generic i32, *only* the value defined
476 by a ``gc.statepoint`` is legal here.
481 The ``gc.result`` represents the return value of the call target of
482 the ``statepoint``. The type of the ``gc.result`` must exactly match
483 the type of the target. If the call target returns void, there will
486 A ``gc.result`` is modeled as a 'readnone' pure function. It has no
487 side effects since it is just a projection of the return value of the
488 previous call represented by the ``gc.statepoint``.
490 'llvm.experimental.gc.relocate' Intrinsic
491 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
498 declare <pointer type>
499 @llvm.experimental.gc.relocate(i32 %statepoint_token,
506 A ``gc.relocate`` returns the potentially relocated value of a pointer
512 The first argument is the ``gc.statepoint`` which starts the
513 safepoint sequence of which this ``gc.relocation`` is a part.
514 Despite the typing of this as a generic i32, *only* the value defined
515 by a ``gc.statepoint`` is legal here.
517 The second argument is an index into the statepoints list of arguments
518 which specifies the allocation for the pointer being relocated.
519 This index must land within the 'gc parameter' section of the
520 statepoint's argument list. The associated value must be within the
521 object with which the pointer being relocated is associated. The optimizer
522 is free to change *which* interior derived pointer is reported, provided that
523 it does not replace an actual base pointer with another interior derived
524 pointer. Collectors are allowed to rely on the base pointer operand
525 remaining an actual base pointer if so constructed.
527 The third argument is an index into the statepoint's list of arguments
528 which specify the (potentially) derived pointer being relocated. It
529 is legal for this index to be the same as the second argument
530 if-and-only-if a base pointer is being relocated. This index must land
531 within the 'gc parameter' section of the statepoint's argument list.
536 The return value of ``gc.relocate`` is the potentially relocated value
537 of the pointer specified by it's arguments. It is unspecified how the
538 value of the returned pointer relates to the argument to the
539 ``gc.statepoint`` other than that a) it points to the same source
540 language object with the same offset, and b) the 'based-on'
541 relationship of the newly relocated pointers is a projection of the
542 unrelocated pointers. In particular, the integer value of the pointer
543 returned is unspecified.
545 A ``gc.relocate`` is modeled as a ``readnone`` pure function. It has no
546 side effects since it is just a way to extract information about work
547 done during the actual call modeled by the ``gc.statepoint``.
549 .. _statepoint-stackmap-format:
554 Locations for each pointer value which may need read and/or updated by
555 the runtime or collector are provided via the :ref:`Stack Map format
556 <stackmap-format>` specified in the PatchPoint documentation.
558 Each statepoint generates the following Locations:
560 * Constant which describes the calling convention of the call target. This
561 constant is a valid :ref:`calling convention identifier <callingconv>` for
562 the version of LLVM used to generate the stackmap. No additional compatibility
563 guarantees are made for this constant over what LLVM provides elsewhere w.r.t.
565 * Constant which describes the flags passed to the statepoint intrinsic
566 * Constant which describes number of following deopt *Locations* (not
568 * Variable number of Locations, one for each deopt parameter listed in
569 the IR statepoint (same number as described by previous Constant)
570 * Variable number of Locations pairs, one pair for each unique pointer
571 which needs relocated. The first Location in each pair describes
572 the base pointer for the object. The second is the derived pointer
573 actually being relocated. It is guaranteed that the base pointer
574 must also appear explicitly as a relocation pair if used after the
575 statepoint. There may be fewer pairs then gc parameters in the IR
576 statepoint. Each *unique* pair will occur at least once; duplicates
579 Note that the Locations used in each section may describe the same
580 physical location. e.g. A stack slot may appear as a deopt location,
581 a gc base pointer, and a gc derived pointer.
583 The LiveOut section of the StkMapRecord will be empty for a statepoint
586 Safepoint Semantics & Verification
587 ==================================
589 The fundamental correctness property for the compiled code's
590 correctness w.r.t. the garbage collector is a dynamic one. It must be
591 the case that there is no dynamic trace such that a operation
592 involving a potentially relocated pointer is observably-after a
593 safepoint which could relocate it. 'observably-after' is this usage
594 means that an outside observer could observe this sequence of events
595 in a way which precludes the operation being performed before the
598 To understand why this 'observable-after' property is required,
599 consider a null comparison performed on the original copy of a
600 relocated pointer. Assuming that control flow follows the safepoint,
601 there is no way to observe externally whether the null comparison is
602 performed before or after the safepoint. (Remember, the original
603 Value is unmodified by the safepoint.) The compiler is free to make
604 either scheduling choice.
606 The actual correctness property implemented is slightly stronger than
607 this. We require that there be no *static path* on which a
608 potentially relocated pointer is 'observably-after' it may have been
609 relocated. This is slightly stronger than is strictly necessary (and
610 thus may disallow some otherwise valid programs), but greatly
611 simplifies reasoning about correctness of the compiled code.
613 By construction, this property will be upheld by the optimizer if
614 correctly established in the source IR. This is a key invariant of
617 The existing IR Verifier pass has been extended to check most of the
618 local restrictions on the intrinsics mentioned in their respective
619 documentation. The current implementation in LLVM does not check the
620 key relocation invariant, but this is ongoing work on developing such
621 a verifier. Please ask on llvm-dev if you're interested in
622 experimenting with the current version.
624 .. _statepoint-utilities:
626 Utility Passes for Safepoint Insertion
627 ======================================
629 .. _RewriteStatepointsForGC:
631 RewriteStatepointsForGC
632 ^^^^^^^^^^^^^^^^^^^^^^^^
634 The pass RewriteStatepointsForGC transforms a functions IR by replacing a
635 ``gc.statepoint`` (with an optional ``gc.result``) with a full relocation
636 sequence, including all required ``gc.relocates``. To function, the pass
637 requires that the GC strategy specified for the function be able to reliably
638 distinguish between GC references and non-GC references in IR it is given.
640 As an example, given this code:
644 define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj)
645 gc "statepoint-example" {
646 call i32 (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 5, i32 0, i32 -1, i32 0, i32 0, i32 0)
647 ret i8 addrspace(1)* %obj
650 The pass would produce this IR:
654 define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj)
655 gc "statepoint-example" {
656 %0 = call i32 (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 5, i32 0, i32 -1, i32 0, i32 0, i32 0, i8 addrspace(1)* %obj)
657 %obj.relocated = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(i32 %0, i32 12, i32 12)
658 ret i8 addrspace(1)* %obj.relocated
661 In the above examples, the addrspace(1) marker on the pointers is the mechanism
662 that the ``statepoint-example`` GC strategy uses to distinguish references from
663 non references. Address space 1 is not globally reserved for this purpose.
665 This pass can be used an utility function by a language frontend that doesn't
666 want to manually reason about liveness, base pointers, or relocation when
667 constructing IR. As currently implemented, RewriteStatepointsForGC must be
668 run after SSA construction (i.e. mem2ref).
670 RewriteStatepointsForGC will ensure that appropriate base pointers are listed
671 for every relocation created. It will do so by duplicating code as needed to
672 propagate the base pointer associated with each pointer being relocated to
673 the appropriate safepoints. The implementation assumes that the following
674 IR constructs produce base pointers: loads from the heap, addresses of global
675 variables, function arguments, function return values. Constant pointers (such
676 as null) are also assumed to be base pointers. In practice, this constraint
677 can be relaxed to producing interior derived pointers provided the target
678 collector can find the associated allocation from an arbitrary interior
681 In practice, RewriteStatepointsForGC can be run much later in the pass
682 pipeline, after most optimization is already done. This helps to improve
683 the quality of the generated code when compiled with garbage collection support.
684 In the long run, this is the intended usage model. At this time, a few details
685 have yet to be worked out about the semantic model required to guarantee this
686 is always correct. As such, please use with caution and report bugs.
693 The pass PlaceSafepoints transforms a function's IR by replacing any call or
694 invoke instructions with appropriate ``gc.statepoint`` and ``gc.result`` pairs,
695 and inserting safepoint polls sufficient to ensure running code checks for a
696 safepoint request on a timely manner. This pass is expected to be run before
697 RewriteStatepointsForGC and thus does not produce full relocation sequences.
699 As an example, given input IR of the following:
703 define void @test() gc "statepoint-example" {
708 declare void @do_safepoint()
709 define void @gc.safepoint_poll() {
710 call void @do_safepoint()
715 This pass would produce the following IR:
719 define void @test() gc "statepoint-example" {
720 %safepoint_token = call i32 (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @do_safepoint, i32 0, i32 0, i32 0, i32 0)
721 %safepoint_token1 = call i32 (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 0)
725 In this case, we've added an (unconditional) entry safepoint poll and converted the call into a ``gc.statepoint``. Note that despite appearances, the entry poll is not necessarily redundant. We'd have to know that ``foo`` and ``test`` were not mutually recursive for the poll to be redundant. In practice, you'd probably want to your poll definition to contain a conditional branch of some form.
728 At the moment, PlaceSafepoints can insert safepoint polls at method entry and
729 loop backedges locations. Extending this to work with return polls would be
730 straight forward if desired.
732 PlaceSafepoints includes a number of optimizations to avoid placing safepoint
733 polls at particular sites unless needed to ensure timely execution of a poll
734 under normal conditions. PlaceSafepoints does not attempt to ensure timely
735 execution of a poll under worst case conditions such as heavy system paging.
737 The implementation of a safepoint poll action is specified by looking up a
738 function of the name ``gc.safepoint_poll`` in the containing Module. The body
739 of this function is inserted at each poll site desired. While calls or invokes
740 inside this method are transformed to a ``gc.statepoints``, recursive poll
741 insertion is not performed.
743 By default PlaceSafepoints passes in ``0xABCDEF00`` as the statepoint
744 ID and ``0`` as the number of patchable bytes to the newly constructed
745 ``gc.statepoint``. These values can be configured on a per-callsite
746 basis using the attributes ``"statepoint-id"`` and
747 ``"statepoint-num-patch-bytes"``. If a call site is marked with a
748 ``"statepoint-id"`` function attribute and its value is a positive
749 integer (represented as a string), then that value is used as the ID
750 of the newly constructed ``gc.statepoint``. If a call site is marked
751 with a ``"statepoint-num-patch-bytes"`` function attribute and its
752 value is a positive integer, then that value is used as the 'num patch
753 bytes' parameter of the newly constructed ``gc.statepoint``. The
754 ``"statepoint-id"`` and ``"statepoint-num-patch-bytes"`` attributes
755 are not propagated to the ``gc.statepoint`` call or invoke if they
756 could be successfully parsed.
758 If you are scheduling the RewriteStatepointsForGC pass late in the pass order,
759 you should probably schedule this pass immediately before it. The exception
760 would be if you need to preserve abstract frame information (e.g. for
761 deoptimization or introspection) at safepoints. In that case, ask on the
762 llvm-dev mailing list for suggestions.
765 Supported Architectures
766 =======================
768 Support for statepoint generation requires some code for each backend.
769 Today, only X86_64 is supported.
771 Bugs and Enhancements
772 =====================
774 Currently known bugs and enhancements under consideration can be
775 tracked by performing a `bugzilla search
776 <http://llvm.org/bugs/buglist.cgi?cmdtype=runnamed&namedcmd=Statepoint%20Bugs&list_id=64342>`_
777 for [Statepoint] in the summary field. When filing new bugs, please
778 use this tag so that interested parties see the newly filed bug. As
779 with most LLVM features, design discussions take place on `llvm-dev
780 <http://lists.llvm.org/mailman/listinfo/llvm-dev>`_, and patches
781 should be sent to `llvm-commits
782 <http://lists.llvm.org/mailman/listinfo/llvm-commits>`_ for review.