| blob | 6663f0caa75fb1d7f1a4b49ca24c8225872677fa |
3 <html>
4 <head>
7 </head>
8 <body>
11 LLVM Alias Analysis Infrastructure
12 </div>
14 <ol>
18 <ul>
23 </ul>
24 </li>
27 <ul>
34 </ul>
35 </li>
38 <ul>
42 </ul>
43 </li>
46 <ul>
50 implementations</a></li>
51 </ul>
52 </li>
54 </ol>
58 </div>
60 <!-- *********************************************************************** -->
63 </div>
64 <!-- *********************************************************************** -->
68 <p>Alias Analysis (aka Pointer Analysis) is a class of techniques which attempt
69 to determine whether or not two pointers ever can point to the same object in
70 memory. There are many different algorithms for alias analysis and many
71 different ways of classifying them: flow-sensitive vs flow-insensitive,
72 context-sensitive vs context-insensitive, field-sensitive vs field-insensitive,
73 unification-based vs subset-based, etc. Traditionally, alias analyses respond
75 indicating that two pointers always point to the same object, might point to the
76 same object, or are known to never point to the same object.</p>
80 class is the primary interface used by clients and implementations of alias
81 analyses in the LLVM system. This class is the common interface between clients
82 of alias analysis information and the implementations providing it, and is
83 designed to support a wide range of implementations and clients (but currently
84 all clients are assumed to be flow-insensitive). In addition to simple alias
85 analysis information, this class exposes Mod/Ref information from those
86 implementations which can provide it, allowing for powerful analyses and
87 transformations to work well together.</p>
89 <p>This document contains information necessary to successfully implement this
90 interface, use it, and to test both sides. It also explains some of the finer
91 points about what exactly results mean. If you feel that something is unclear
95 </div>
97 <!-- *********************************************************************** -->
100 </div>
101 <!-- *********************************************************************** -->
107 class defines the interface that the various alias analysis implementations
110 mod/ref query, respectively.</p>
113 represented in several different ways. In particular, memory objects are
114 represented as a starting address and size, and function calls are represented
117 which allow you to get mod/ref information for arbitrary instructions.</p>
119 </div>
121 <!-- ======================================================================= -->
124 </div>
129 which are used to query whether or not two memory objects alias, whether
130 function calls can modify or read a memory object, etc. For all of these
131 queries, memory objects are represented as a pair of their starting address (a
134 <p>Representing memory objects as a starting address and a size is critically
135 important for correct Alias Analyses. For example, consider this (silly, but
136 possible) C code:</p>
139 <pre>
140 int i;
141 char C[2];
142 char A[10];
143 /* ... */
145 C[0] = A[i]; /* One byte store */
147 }
148 </pre>
149 </div>
153 locations one byte apart, and the accesses are each one byte. In this case, the
154 LICM pass can use store motion to remove the stores from the loop. In
155 constrast, the following code:</p>
158 <pre>
159 int i;
160 char C[2];
161 char A[10];
162 /* ... */
164 ((short*)C)[0] = A[i]; /* Two byte store! */
166 }
167 </pre>
168 </div>
170 <p>In this case, the two stores to C do alias each other, because the access to
172 available in the query, even the first case would have to conservatively assume
173 that the accesses alias.</p>
175 </div>
177 <!-- ======================================================================= -->
180 </div>
184 not two memory objects alias each other. It takes two memory objects as input
185 and returns MustAlias, MayAlias, or NoAlias as appropriate.
186 </div>
188 <!-- _______________________________________________________________________ -->
191 </div>
195 <p>An Alias Analysis implementation can return one of three responses:
196 MustAlias, MayAlias, and NoAlias. The No and May alias results are obvious: if
197 the two pointers can never equal each other, return NoAlias, if they might,
198 return MayAlias.</p>
200 <p>The MustAlias response is trickier though. In LLVM, the Must Alias response
201 may only be returned if the two memory objects are guaranteed to always start at
202 exactly the same location. If two memory objects overlap, but do not start at
203 the same location, return MayAlias.</p>
205 </div>
207 <!-- ======================================================================= -->
210 </div>
215 execution of an instruction can read or modify a memory location. Mod/Ref
217 a location, ModRef is returned.</p>
220 method for testing dependencies between function calls. This method takes two
221 call sites (CS1 & CS2), returns NoModRef if the two calls refer to disjoint
222 memory locations, Ref if CS1 reads memory written by CS2, Mod if CS1 writes to
223 memory read or written by CS2, or ModRef if CS1 might read or write memory
224 accessed by CS2. Note that this relation is not commutative. Clients that use
226 method, which indicates whether or not an analysis can provide mod/ref
227 information for function call pairs (most can not). If this predicate is false,
229 method many times.</p>
231 </div>
234 <!-- ======================================================================= -->
237 </div>
241 <p>
242 Several other tidbits of information are often collected by various alias
243 analysis implementations and can be put to good use by various clients.
244 </p>
246 </div>
248 <!-- _______________________________________________________________________ -->
251 </div>
256 always must alias a pointer. This information can be provided in some cases for
257 important objects like the null pointer and global values. Knowing that a
258 pointer always points to a particular function allows indirect calls to be
259 turned into direct calls, for example.</p>
261 </div>
263 <!-- _______________________________________________________________________ -->
266 </div>
271 analysis can prove that the pointer only points to unchanging memory locations
272 (functions, constant global variables, and the null pointer). This information
273 can be used to refine mod/ref information: it is impossible for an unchanging
274 memory location to be modified.</p>
276 </div>
278 <!-- _______________________________________________________________________ -->
282 </div>
286 <p>These methods are used to provide very simple mod/ref information for
288 function if the analysis can prove that the function never reads or writes to
289 memory, or if the function only reads from constant memory. Functions with this
290 property are side-effect free and only depend on their input arguments, allowing
291 them to be eliminated if they form common subexpressions or be hoisted out of
297 can prove that (at most) the function only reads from non-volatile memory.
298 Functions with this property are side-effect free, only depending on their input
299 arguments and the state of memory when they are called. This property allows
300 calls to these functions to be eliminated and moved around, as long as there is
301 no store instruction that changes the contents of memory. Note that all
305 </div>
307 <!-- *********************************************************************** -->
310 </div>
311 <!-- *********************************************************************** -->
315 <p>Writing a new alias analysis implementation for LLVM is quite
316 straight-forward. There are already several implementations that you can use
317 for examples, and the following information should help fill in any details.
321 </div>
323 <!-- ======================================================================= -->
326 </div>
332 Analysis. As is the case with most other analyses and transformations, the
333 answer should be fairly obvious from what type of problem you are trying to
334 solve:</p>
336 <ol>
337 <li>If you require interprocedural analysis, it should be a
340 <li>If you don't need to look at the program at all, subclass
342 </ol>
344 <p>In addition to the pass that you subclass, you should also inherit from the
349 </div>
351 <!-- ======================================================================= -->
354 </div>
363 declaring any pass dependencies your pass has. Thus you should have something
364 like this:</p>
367 <pre>
368 void getAnalysisUsage(AnalysisUsage &AU) const {
369 AliasAnalysis::getAnalysisUsage(AU);
371 }
372 </pre>
373 </div>
381 <pre>
382 bool run(Module &M) {
383 InitializeAliasAnalysis(this);
385 return false;
386 }
387 </pre>
388 </div>
390 </div>
392 <!-- ======================================================================= -->
395 </div>
402 alias analysis implementation, which ends up returning conservatively correct
404 respectively). Depending on the capabilities of the analysis you are
405 implementing, you just override the interfaces you can improve.</p>
407 </div>
411 <!-- ======================================================================= -->
414 </div>
420 passes) every alias analysis pass chains to another alias analysis
421 implementation (for example, the user can specify "<tt>-basicaa -ds-aa
422 -anders-aa -licm</tt>" to get the maximum benefit from the three alias
423 analyses). The alias analysis class automatically takes care of most of this
424 for methods that you don't override. For methods that you do override, in code
425 paths that return a conservative MayAlias or Mod/Ref result, simply return
426 whatever the superclass computes. For example:</p>
429 <pre>
430 AliasAnalysis::AliasResult alias(const Value *V1, unsigned V1Size,
431 const Value *V2, unsigned V2Size) {
432 if (...)
433 return NoAlias;
434 ...
437 return AliasAnalysis::alias(V1, V1Size, V2, V2Size);
438 }
439 </pre>
440 </div>
442 <p>In addition to analysis queries, you must make sure to unconditionally pass
444 well if you override them, which allows all alias analyses in a change to be
445 updated.</p>
447 </div>
450 <!-- ======================================================================= -->
453 </div>
456 <p>
457 Alias analysis information is initially computed for a static snapshot of the
458 program, but clients will use this information to make transformations to the
459 code. All but the most trivial forms of alias analysis will need to have their
460 analysis results updated to reflect the changes made by these transformations.
461 </p>
463 <p>
465 communicate program changes from the clients to the analysis implementations.
466 Various alias analysis implementations should use these methods to ensure that
467 their internal data structures are kept up-to-date as the program changes (for
468 example, when an instruction is deleted), and clients of alias analysis must be
469 sure to call these interfaces appropriately.
470 </p>
471 </div>
473 <!-- _______________________________________________________________________ -->
478 instruction or any other value from the program (including values that do not
479 use pointers). Typically alias analyses keep data structures that have entries
480 for each value in the program. When this method is called, they should remove
481 any entries for the specified value, if they exist.
482 </div>
484 <!-- _______________________________________________________________________ -->
489 program. There is no way to introduce a value into the program that did not
490 exist before (this doesn't make sense for a safe compiler transformation), so
491 this is the only way to introduce a new value. This method indicates that the
492 new value has exactly the same properties as the value being copied.
493 </div>
495 <!-- _______________________________________________________________________ -->
499 This method is a simple helper method that is provided to make clients easier to
500 use. It is implemented by copying the old analysis information to the new
501 value, then deleting the old value. This method cannot be overridden by alias
502 analysis implementations.
503 </div>
505 <!-- ======================================================================= -->
508 </div>
512 <p>From the LLVM perspective, the only thing you need to do to provide an
514 serviced quickly. The actual calculation of the alias analysis results (the
515 "run" method) is only performed once, but many (perhaps duplicate) queries may
516 be performed. Because of this, try to move as much computation to the run
517 method as possible (within reason).</p>
519 </div>
521 <!-- *********************************************************************** -->
524 </div>
525 <!-- *********************************************************************** -->
529 <p>There are several different ways to use alias analysis results. In order of
530 preference, these are...</p>
532 </div>
534 <!-- ======================================================================= -->
537 </div>
543 or transformation can be modeled in a form that uses value numbering
544 information, you don't have to do anything special to handle load instructions:
547 </div>
549 <!-- ======================================================================= -->
552 </div>
557 in some scope, rather than information about pairwise aliasing. The <tt><a
559 is used to efficiently build these Alias Sets from the pairwise alias analysis
563 to add information about various potentially aliasing instructions in the scope
564 you are interested in. Once all of the alias sets are completed, your pass
565 should simply iterate through the constructed alias sets, using the
569 to be disjoint, calculate mod/ref information and volatility for the set, and
570 keep track of whether or not all of the pointers in the set are Must aliases.
571 The AliasSetTracker also makes sure that sets are properly folded due to call
572 instructions, and can provide a list of pointers in each set.</p>
577 then all load instructions from that set may be hoisted out of the loop. If any
579 sunk to outside of the loop, promoting the memory location to a register for the
580 duration of the loop nest. Both of these transformations only apply if the
581 pointer argument is loop-invariant.</p>
583 </div>
585 <!-- _______________________________________________________________________ -->
587 The AliasSetTracker implementation
588 </div>
592 <p>The AliasSetTracker class is implemented to be as efficient as possible. It
593 uses the union-find algorithm to efficiently merge AliasSets when a pointer is
594 inserted into the AliasSetTracker that aliases multiple sets. The primary data
595 structure is a hash table mapping pointers to the AliasSet they are in.</p>
597 <p>The AliasSetTracker class must maintain a list of all of the LLVM Value*'s
598 that are in each AliasSet. Since the hash table already has entries for each
599 LLVM Value* of interest, the AliasesSets thread the linked list through these
600 hash-table nodes to avoid having to allocate memory unnecessarily, and to make
601 merging alias sets extremely efficient (the linked list merge is constant time).
602 </p>
604 <p>You shouldn't need to understand these details if you are just a client of
605 the AliasSetTracker, but if you look at the code, hopefully this brief
606 description will help make sense of why things are designed the way they
607 are.</p>
609 </div>
611 <!-- ======================================================================= -->
614 </div>
618 <p>If neither of these utility class are what your pass needs, you should use
620 the higher-level methods when possible (e.g., use mod/ref information instead of
622 best precision and efficiency.</p>
624 </div>
626 <!-- *********************************************************************** -->
629 </div>
630 <!-- *********************************************************************** -->
634 <p>If you're going to be working with the LLVM alias analysis infrastructure,
635 you should know what clients and implementations of alias analysis are
636 available. In particular, if you are implementing an alias analysis, you should
638 for monitoring and evaluating different implementations.</p>
640 </div>
642 <!-- ======================================================================= -->
645 </div>
654 </div>
656 <!-- _______________________________________________________________________ -->
659 </div>
664 never returns any useful information. This pass can be useful if you think that
665 alias analysis is doing something wrong and are trying to narrow down a
666 problem.</p>
668 </div>
670 <!-- _______________________________________________________________________ -->
673 </div>
680 <ul>
681 <li>Distinct globals, stack allocations, and heap allocations can never
682 alias.</li>
683 <li>Globals, stack allocations, and heap allocations never alias the null
684 pointer.</li>
689 <li>Pointers that obviously point to constant globals
691 <li>Function calls can not modify or references stack allocations if they never
692 escape from the function that allocates them (a common case for automatic
693 arrays).</li>
694 </ul>
696 </div>
698 <!-- _______________________________________________________________________ -->
701 </div>
705 <p>This pass implements a simple context-sensitive mod/ref and alias analysis
706 for internal global variables that don't "have their address taken". If a
707 global does not have its address taken, the pass knows that no pointers alias
708 the global. This pass also keeps track of functions that it knows never access
709 memory or never read memory. This allows certain optimizations (e.g. GCSE) to
710 eliminate call instructions entirely.
711 </p>
713 <p>The real power of this pass is that it provides context-sensitive mod/ref
714 information for call instructions. This allows the optimizer to know that
715 calls to a function do not clobber or read the value of the global, allowing
716 loads and stores to be eliminated.</p>
718 <p>Note that this pass is somewhat limited in its scope (only support
719 non-address taken globals), but is very quick analysis.</p>
720 </div>
722 <!-- _______________________________________________________________________ -->
725 </div>
730 for interprocedural alias analysis. This algorithm is a subset-based,
731 flow-insensitive, context-insensitive, and field-insensitive alias analysis that
732 is widely believed to be fairly precise. Unfortunately, this algorithm is also
735 substitution") to improve its efficiency, so it can be quite slow in common
736 cases.
737 </p>
739 </div>
741 <!-- _______________________________________________________________________ -->
744 </div>
749 "Steensgaard's algorithm" for interprocedural alias analysis. Steensgaard's
750 algorithm is a unification-based, flow-insensitive, context-insensitive, and
751 field-insensitive alias analysis that is also very scalable (effectively linear
752 time).</p>
755 field-<b>sensitive</b>" version of Steensgaard's algorithm using the Data
756 Structure Analysis framework. This gives it substantially more precision than
757 the standard algorithm while maintaining excellent analysis scalability.</p>
760 module, it is not part of the LLVM core.</p>
762 </div>
764 <!-- _______________________________________________________________________ -->
767 </div>
772 algorithm. Data Structure Analysis is a modular unification-based,
775 O(n*log(n)).</p>
777 <p>This algorithm is capable of responding to a full variety of alias analysis
778 queries, and can provide context-sensitive mod/ref information as well. The
779 only major facility not implemented so far is support for must-alias
780 information.</p>
783 module, it is not part of the LLVM core.</p>
785 </div>
788 <!-- ======================================================================= -->
791 </div>
794 LLVM includes several alias-analysis driven transformations which can be used
795 with any of the implementations above.
796 </div>
798 <!-- _______________________________________________________________________ -->
801 </div>
807 not have side-effects and are not used.</p>
809 </div>
812 <!-- _______________________________________________________________________ -->
815 </div>
821 different transformations:</p>
823 <ul>
824 <li>It uses mod/ref information to hoist or sink load instructions out of loops
825 if there are no instructions in the loop that modifies the memory loaded.</li>
827 <li>It uses mod/ref information to hoist function calls out of loops that do not
828 write to memory and are loop-invariant.</li>
830 <li>If uses alias information to promote memory objects that are loaded and
831 stored to in loops to live in a register instead. It can do this if there are
832 no may aliases to the loaded/stored memory location.</li>
833 </ul>
835 </div>
837 <!-- _______________________________________________________________________ -->
840 </div>
843 <p>
845 by-value instead. In particular, if pointer arguments are only loaded from it
846 passes in the value loaded instead of the address to the function. This pass
847 uses alias information to make sure that the value loaded from the argument
848 pointer is not modified between the entry of the function and any load of the
849 pointer.</p>
850 </div>
852 <!-- _______________________________________________________________________ -->
855 </div>
860 number</a>" loads and pointers values, which is used by the GCSE pass to
862 and must-alias information. This combination of passes can make the following
863 transformations:</p>
865 <ul>
867 <li>Load instructions that follow a store to the same location are replaced with
869 <li>Pointers values (e.g. formal arguments) that must-alias simpler expressions
870 (e.g. global variables or the null pointer) are replaced. Note that this
871 implements transformations like "virtual method resolution", turning indirect
872 calls into direct calls.</li>
873 </ul>
875 </div>
877 <!-- ======================================================================= -->
880 implementations</a>
881 </div>
885 <p>These passes are useful for evaluating the various alias analysis
886 implementations. You can use them with commands like '<tt>opt -anders-aa -ds-aa
889 </div>
891 <!-- _______________________________________________________________________ -->
894 </div>
904 <pre>
905 % opt -ds-aa -print-alias-sets -disable-output
906 </pre>
907 </div>
909 </div>
912 <!-- _______________________________________________________________________ -->
915 </div>
920 pass is making and what responses are returned by the alias analysis. As an
921 example,</p>
924 <pre>
925 % opt -basicaa -count-aa -ds-aa -count-aa -licm
926 </pre>
927 </div>
929 <p>will print out how many queries (and what responses are returned) by the
932 when debugging a transformation or an alias analysis implementation.</p>
934 </div>
936 <!-- _______________________________________________________________________ -->
939 </div>
944 function and asks an alias analysis whether or not the pointers alias. This
945 gives an indication of the precision of the alias analysis. Statistics are
946 printed indicating the percent of no/may/must aliases found (a more precise
947 algorithm will have a lower number of may aliases).</p>
949 </div>
951 <!-- *********************************************************************** -->
954 </div>
955 <!-- *********************************************************************** -->
959 <p>If you're just looking to be a client of alias analysis information, consider
960 using the Memory Dependence Analysis interface instead. MemDep is a lazy,
961 caching layer on top of alias analysis that is able to answer the question of
962 what preceding memory operations a given instruction depends on, either at an
963 intra- or inter-block level. Because of its laziness and caching
964 policy, using MemDep can be a significant performance win over accessing alias
965 analysis directly.</p>
967 </div>
969 <!-- *********************************************************************** -->
971 <hr>
972 <address>
980 Last modified: $Date$
981 </address>
983 </body>
984 </html>