This file is indexed.

/usr/share/doc/llvm-3.4-doc/html/GarbageCollection.html is in llvm-3.4-doc 1:3.4-1ubuntu3.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39
  40
  41
  42
  43
  44
  45
  46
  47
  48
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
  "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">


<html xmlns="http://www.w3.org/1999/xhtml">
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
    
    <title>Accurate Garbage Collection with LLVM &mdash; LLVM 3.4 documentation</title>
    
    <link rel="stylesheet" href="_static/llvm-theme.css" type="text/css" />
    <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
    
    <script type="text/javascript">
      var DOCUMENTATION_OPTIONS = {
        URL_ROOT:    './',
        VERSION:     '3.4',
        COLLAPSE_INDEX: false,
        FILE_SUFFIX: '.html',
        HAS_SOURCE:  true
      };
    </script>
    <script type="text/javascript" src="_static/jquery.js"></script>
    <script type="text/javascript" src="_static/underscore.js"></script>
    <script type="text/javascript" src="_static/doctools.js"></script>
    <link rel="top" title="LLVM 3.4 documentation" href="index.html" />
    <link rel="next" title="Writing an LLVM Pass" href="WritingAnLLVMPass.html" />
    <link rel="prev" title="How To Use Instruction Mappings" href="HowToUseInstrMappings.html" />
<style type="text/css">
  table.right { float: right; margin-left: 20px; }
  table.right td { border: 1px solid #ccc; }
</style>

  </head>
  <body>
<div class="logo">
  <a href="index.html">
    <img src="_static/logo.png"
         alt="LLVM Logo" width="250" height="88"/></a>
</div>

    <div class="related">
      <h3>Navigation</h3>
      <ul>
        <li class="right" style="margin-right: 10px">
          <a href="genindex.html" title="General Index"
             accesskey="I">index</a></li>
        <li class="right" >
          <a href="WritingAnLLVMPass.html" title="Writing an LLVM Pass"
             accesskey="N">next</a> |</li>
        <li class="right" >
          <a href="HowToUseInstrMappings.html" title="How To Use Instruction Mappings"
             accesskey="P">previous</a> |</li>
  <li><a href="http://llvm.org/">LLVM Home</a>&nbsp;|&nbsp;</li>
  <li><a href="index.html">Documentation</a>&raquo;</li>
 
      </ul>
    </div>


    <div class="document">
      <div class="documentwrapper">
          <div class="body">
            
  <div class="section" id="accurate-garbage-collection-with-llvm">
<h1>Accurate Garbage Collection with LLVM<a class="headerlink" href="#accurate-garbage-collection-with-llvm" title="Permalink to this headline"></a></h1>
<div class="contents local topic" id="contents">
<ul class="simple">
<li><a class="reference internal" href="#introduction" id="id1">Introduction</a><ul>
<li><a class="reference internal" href="#goals-and-non-goals" id="id2">Goals and non-goals</a></li>
</ul>
</li>
<li><a class="reference internal" href="#getting-started" id="id3">Getting started</a><ul>
<li><a class="reference internal" href="#in-your-compiler" id="id4">In your compiler</a></li>
<li><a class="reference internal" href="#in-your-runtime" id="id5">In your runtime</a></li>
<li><a class="reference internal" href="#about-the-shadow-stack" id="id6">About the shadow stack</a></li>
</ul>
</li>
<li><a class="reference internal" href="#ir-features" id="id7">IR features</a><ul>
<li><a class="reference internal" href="#specifying-gc-code-generation-gc" id="id8">Specifying GC code generation: <tt class="docutils literal"><span class="pre">gc</span> <span class="pre">&quot;...&quot;</span></tt></a></li>
<li><a class="reference internal" href="#gcroot" id="id9">Identifying GC roots on the stack: <tt class="docutils literal"><span class="pre">llvm.gcroot</span></tt></a></li>
<li><a class="reference internal" href="#reading-and-writing-references-in-the-heap" id="id10">Reading and writing references in the heap</a><ul>
<li><a class="reference internal" href="#write-barrier-llvm-gcwrite" id="id11">Write barrier: <tt class="docutils literal"><span class="pre">llvm.gcwrite</span></tt></a></li>
<li><a class="reference internal" href="#read-barrier-llvm-gcread" id="id12">Read barrier: <tt class="docutils literal"><span class="pre">llvm.gcread</span></tt></a></li>
</ul>
</li>
</ul>
</li>
<li><a class="reference internal" href="#implementing-a-collector-plugin" id="id13">Implementing a collector plugin</a><ul>
<li><a class="reference internal" href="#overview-of-available-features" id="id14">Overview of available features</a></li>
<li><a class="reference internal" href="#computing-stack-maps" id="id15">Computing stack maps</a></li>
<li><a class="reference internal" href="#initializing-roots-to-null-initroots" id="id16">Initializing roots to null: <tt class="docutils literal"><span class="pre">InitRoots</span></tt></a></li>
<li><a class="reference internal" href="#custom-lowering-of-intrinsics-customroots-customreadbarriers-and-customwritebarriers" id="id17">Custom lowering of intrinsics: <tt class="docutils literal"><span class="pre">CustomRoots</span></tt>, <tt class="docutils literal"><span class="pre">CustomReadBarriers</span></tt>, and <tt class="docutils literal"><span class="pre">CustomWriteBarriers</span></tt></a></li>
<li><a class="reference internal" href="#generating-safe-points-neededsafepoints" id="id18">Generating safe points: <tt class="docutils literal"><span class="pre">NeededSafePoints</span></tt></a></li>
<li><a class="reference internal" href="#emitting-assembly-code-gcmetadataprinter" id="id19">Emitting assembly code: <tt class="docutils literal"><span class="pre">GCMetadataPrinter</span></tt></a></li>
</ul>
</li>
<li><a class="reference internal" href="#references" id="id20">References</a></li>
</ul>
</div>
<div class="section" id="introduction">
<h2><a class="toc-backref" href="#id1">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline"></a></h2>
<p>Garbage collection is a widely used technique that frees the programmer from
having to know the lifetimes of heap objects, making software easier to produce
and maintain.  Many programming languages rely on garbage collection for
automatic memory management.  There are two primary forms of garbage collection:
conservative and accurate.</p>
<p>Conservative garbage collection often does not require any special support from
either the language or the compiler: it can handle non-type-safe programming
languages (such as C/C++) and does not require any special information from the
compiler.  The <a class="reference external" href="http://www.hpl.hp.com/personal/Hans_Boehm/gc/">Boehm collector</a> is an example of a
state-of-the-art conservative collector.</p>
<p>Accurate garbage collection requires the ability to identify all pointers in the
program at run-time (which requires that the source-language be type-safe in
most cases).  Identifying pointers at run-time requires compiler support to
locate all places that hold live pointer variables at run-time, including the
<a class="reference internal" href="#gcroot"><em>processor stack and registers</em></a>.</p>
<p>Conservative garbage collection is attractive because it does not require any
special compiler support, but it does have problems.  In particular, because the
conservative garbage collector cannot <em>know</em> that a particular word in the
machine is a pointer, it cannot move live objects in the heap (preventing the
use of compacting and generational GC algorithms) and it can occasionally suffer
from memory leaks due to integer values that happen to point to objects in the
program.  In addition, some aggressive compiler transformations can break
conservative garbage collectors (though these seem rare in practice).</p>
<p>Accurate garbage collectors do not suffer from any of these problems, but they
can suffer from degraded scalar optimization of the program.  In particular,
because the runtime must be able to identify and update all pointers active in
the program, some optimizations are less effective.  In practice, however, the
locality and performance benefits of using aggressive garbage collection
techniques dominates any low-level losses.</p>
<p>This document describes the mechanisms and interfaces provided by LLVM to
support accurate garbage collection.</p>
<div class="section" id="goals-and-non-goals">
<h3><a class="toc-backref" href="#id2">Goals and non-goals</a><a class="headerlink" href="#goals-and-non-goals" title="Permalink to this headline"></a></h3>
<p>LLVM&#8217;s intermediate representation provides <a class="reference internal" href="#gc-intrinsics"><em>garbage collection intrinsics</em></a> that offer support for a broad class of collector models.  For
instance, the intrinsics permit:</p>
<ul class="simple">
<li>semi-space collectors</li>
<li>mark-sweep collectors</li>
<li>generational collectors</li>
<li>reference counting</li>
<li>incremental collectors</li>
<li>concurrent collectors</li>
<li>cooperative collectors</li>
</ul>
<p>We hope that the primitive support built into the LLVM IR is sufficient to
support a broad class of garbage collected languages including Scheme, ML, Java,
C#, Perl, Python, Lua, Ruby, other scripting languages, and more.</p>
<p>However, LLVM does not itself provide a garbage collector &#8212; this should be
part of your language&#8217;s runtime library.  LLVM provides a framework for compile
time <a class="reference internal" href="#plugin"><em>code generation plugins</em></a>.  The role of these plugins is to
generate code and data structures which conforms to the <em>binary interface</em>
specified by the <em>runtime library</em>.  This is similar to the relationship between
LLVM and DWARF debugging info, for example.  The difference primarily lies in
the lack of an established standard in the domain of garbage collection &#8212; thus
the plugins.</p>
<p>The aspects of the binary interface with which LLVM&#8217;s GC support is
concerned are:</p>
<ul class="simple">
<li>Creation of GC-safe points within code where collection is allowed to execute
safely.</li>
<li>Computation of the stack map.  For each safe point in the code, object
references within the stack frame must be identified so that the collector may
traverse and perhaps update them.</li>
<li>Write barriers when storing object references to the heap.  These are commonly
used to optimize incremental scans in generational collectors.</li>
<li>Emission of read barriers when loading object references.  These are useful
for interoperating with concurrent collectors.</li>
</ul>
<p>There are additional areas that LLVM does not directly address:</p>
<ul class="simple">
<li>Registration of global roots with the runtime.</li>
<li>Registration of stack map entries with the runtime.</li>
<li>The functions used by the program to allocate memory, trigger a collection,
etc.</li>
<li>Computation or compilation of type maps, or registration of them with the
runtime.  These are used to crawl the heap for object references.</li>
</ul>
<p>In general, LLVM&#8217;s support for GC does not include features which can be
adequately addressed with other features of the IR and does not specify a
particular binary interface.  On the plus side, this means that you should be
able to integrate LLVM with an existing runtime.  On the other hand, it leaves a
lot of work for the developer of a novel language.  However, it&#8217;s easy to get
started quickly and scale up to a more sophisticated implementation as your
compiler matures.</p>
</div>
</div>
<div class="section" id="getting-started">
<h2><a class="toc-backref" href="#id3">Getting started</a><a class="headerlink" href="#getting-started" title="Permalink to this headline"></a></h2>
<p>Using a GC with LLVM implies many things, for example:</p>
<ul class="simple">
<li>Write a runtime library or find an existing one which implements a GC heap.<ol class="arabic">
<li>Implement a memory allocator.</li>
<li>Design a binary interface for the stack map, used to identify references
within a stack frame on the machine stack.*</li>
<li>Implement a stack crawler to discover functions on the call stack.*</li>
<li>Implement a registry for global roots.</li>
<li>Design a binary interface for type maps, used to identify references
within heap objects.</li>
<li>Implement a collection routine bringing together all of the above.</li>
</ol>
</li>
<li>Emit compatible code from your compiler.<ul>
<li>Initialization in the main function.</li>
<li>Use the <tt class="docutils literal"><span class="pre">gc</span> <span class="pre">&quot;...&quot;</span></tt> attribute to enable GC code generation (or
<tt class="docutils literal"><span class="pre">F.setGC(&quot;...&quot;)</span></tt>).</li>
<li>Use <tt class="docutils literal"><span class="pre">&#64;llvm.gcroot</span></tt> to mark stack roots.</li>
<li>Use <tt class="docutils literal"><span class="pre">&#64;llvm.gcread</span></tt> and/or <tt class="docutils literal"><span class="pre">&#64;llvm.gcwrite</span></tt> to manipulate GC references,
if necessary.</li>
<li>Allocate memory using the GC allocation routine provided by the runtime
library.</li>
<li>Generate type maps according to your runtime&#8217;s binary interface.</li>
</ul>
</li>
<li>Write a compiler plugin to interface LLVM with the runtime library.*<ul>
<li>Lower <tt class="docutils literal"><span class="pre">&#64;llvm.gcread</span></tt> and <tt class="docutils literal"><span class="pre">&#64;llvm.gcwrite</span></tt> to appropriate code
sequences.*</li>
<li>Compile LLVM&#8217;s stack map to the binary form expected by the runtime.</li>
</ul>
</li>
<li>Load the plugin into the compiler.  Use <tt class="docutils literal"><span class="pre">llc</span> <span class="pre">-load</span></tt> or link the plugin
statically with your language&#8217;s compiler.*</li>
<li>Link program executables with the runtime.</li>
</ul>
<p>To help with several of these tasks (those indicated with a *), LLVM includes a
highly portable, built-in ShadowStack code generator.  It is compiled into
<tt class="docutils literal"><span class="pre">llc</span></tt> and works even with the interpreter and C backends.</p>
<div class="section" id="in-your-compiler">
<h3><a class="toc-backref" href="#id4">In your compiler</a><a class="headerlink" href="#in-your-compiler" title="Permalink to this headline"></a></h3>
<p>To turn the shadow stack on for your functions, first call:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="n">F</span><span class="p">.</span><span class="n">setGC</span><span class="p">(</span><span class="s">&quot;shadow-stack&quot;</span><span class="p">);</span>
</pre></div>
</div>
<p>for each function your compiler emits. Since the shadow stack is built into
LLVM, you do not need to load a plugin.</p>
<p>Your compiler must also use <tt class="docutils literal"><span class="pre">&#64;llvm.gcroot</span></tt> as documented.  Don&#8217;t forget to
create a root for each intermediate value that is generated when evaluating an
expression.  In <tt class="docutils literal"><span class="pre">h(f(),</span> <span class="pre">g())</span></tt>, the result of <tt class="docutils literal"><span class="pre">f()</span></tt> could easily be collected
if evaluating <tt class="docutils literal"><span class="pre">g()</span></tt> triggers a collection.</p>
<p>There&#8217;s no need to use <tt class="docutils literal"><span class="pre">&#64;llvm.gcread</span></tt> and <tt class="docutils literal"><span class="pre">&#64;llvm.gcwrite</span></tt> over plain
<tt class="docutils literal"><span class="pre">load</span></tt> and <tt class="docutils literal"><span class="pre">store</span></tt> for now.  You will need them when switching to a more
advanced GC.</p>
</div>
<div class="section" id="in-your-runtime">
<h3><a class="toc-backref" href="#id5">In your runtime</a><a class="headerlink" href="#in-your-runtime" title="Permalink to this headline"></a></h3>
<p>The shadow stack doesn&#8217;t imply a memory allocation algorithm.  A semispace
collector or building atop <tt class="docutils literal"><span class="pre">malloc</span></tt> are great places to start, and can be
implemented with very little code.</p>
<p>When it comes time to collect, however, your runtime needs to traverse the stack
roots, and for this it needs to integrate with the shadow stack.  Luckily, doing
so is very simple. (This code is heavily commented to help you understand the
data structure, but there are only 20 lines of meaningful code.)</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="c1">/// @brief The map for a single function&#39;s stack frame.  One of these is</span>
<span class="c1">///        compiled as constant data into the executable for each function.</span>
<span class="c1">///</span>
<span class="c1">/// Storage of metadata values is elided if the %metadata parameter to</span>
<span class="c1">/// @llvm.gcroot is null.</span>
<span class="k">struct</span> <span class="n">FrameMap</span> <span class="p">{</span>
  <span class="kt">int32_t</span> <span class="n">NumRoots</span><span class="p">;</span>    <span class="c1">//&lt; Number of roots in stack frame.</span>
  <span class="kt">int32_t</span> <span class="n">NumMeta</span><span class="p">;</span>     <span class="c1">//&lt; Number of metadata entries.  May be &lt; NumRoots.</span>
  <span class="k">const</span> <span class="kt">void</span> <span class="o">*</span><span class="n">Meta</span><span class="p">[</span><span class="mi">0</span><span class="p">];</span> <span class="c1">//&lt; Metadata for each root.</span>
<span class="p">};</span>

<span class="c1">/// @brief A link in the dynamic shadow stack.  One of these is embedded in</span>
<span class="c1">///        the stack frame of each function on the call stack.</span>
<span class="k">struct</span> <span class="n">StackEntry</span> <span class="p">{</span>
  <span class="n">StackEntry</span> <span class="o">*</span><span class="n">Next</span><span class="p">;</span>    <span class="c1">//&lt; Link to next stack entry (the caller&#39;s).</span>
  <span class="k">const</span> <span class="n">FrameMap</span> <span class="o">*</span><span class="n">Map</span><span class="p">;</span> <span class="c1">//&lt; Pointer to constant FrameMap.</span>
  <span class="kt">void</span> <span class="o">*</span><span class="n">Roots</span><span class="p">[</span><span class="mi">0</span><span class="p">];</span>      <span class="c1">//&lt; Stack roots (in-place array).</span>
<span class="p">};</span>

<span class="c1">/// @brief The head of the singly-linked list of StackEntries.  Functions push</span>
<span class="c1">///        and pop onto this in their prologue and epilogue.</span>
<span class="c1">///</span>
<span class="c1">/// Since there is only a global list, this technique is not threadsafe.</span>
<span class="n">StackEntry</span> <span class="o">*</span><span class="n">llvm_gc_root_chain</span><span class="p">;</span>

<span class="c1">/// @brief Calls Visitor(root, meta) for each GC root on the stack.</span>
<span class="c1">///        root and meta are exactly the values passed to</span>
<span class="c1">///        @llvm.gcroot.</span>
<span class="c1">///</span>
<span class="c1">/// Visitor could be a function to recursively mark live objects.  Or it</span>
<span class="c1">/// might copy them to another heap or generation.</span>
<span class="c1">///</span>
<span class="c1">/// @param Visitor A function to invoke for every GC root on the stack.</span>
<span class="kt">void</span> <span class="nf">visitGCRoots</span><span class="p">(</span><span class="kt">void</span> <span class="p">(</span><span class="o">*</span><span class="n">Visitor</span><span class="p">)(</span><span class="kt">void</span> <span class="o">**</span><span class="n">Root</span><span class="p">,</span> <span class="k">const</span> <span class="kt">void</span> <span class="o">*</span><span class="n">Meta</span><span class="p">))</span> <span class="p">{</span>
  <span class="k">for</span> <span class="p">(</span><span class="n">StackEntry</span> <span class="o">*</span><span class="n">R</span> <span class="o">=</span> <span class="n">llvm_gc_root_chain</span><span class="p">;</span> <span class="n">R</span><span class="p">;</span> <span class="n">R</span> <span class="o">=</span> <span class="n">R</span><span class="o">-&gt;</span><span class="n">Next</span><span class="p">)</span> <span class="p">{</span>
    <span class="kt">unsigned</span> <span class="n">i</span> <span class="o">=</span> <span class="mi">0</span><span class="p">;</span>

    <span class="c1">// For roots [0, NumMeta), the metadata pointer is in the FrameMap.</span>
    <span class="k">for</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="n">e</span> <span class="o">=</span> <span class="n">R</span><span class="o">-&gt;</span><span class="n">Map</span><span class="o">-&gt;</span><span class="n">NumMeta</span><span class="p">;</span> <span class="n">i</span> <span class="o">!=</span> <span class="n">e</span><span class="p">;</span> <span class="o">++</span><span class="n">i</span><span class="p">)</span>
      <span class="n">Visitor</span><span class="p">(</span><span class="o">&amp;</span><span class="n">R</span><span class="o">-&gt;</span><span class="n">Roots</span><span class="p">[</span><span class="n">i</span><span class="p">],</span> <span class="n">R</span><span class="o">-&gt;</span><span class="n">Map</span><span class="o">-&gt;</span><span class="n">Meta</span><span class="p">[</span><span class="n">i</span><span class="p">]);</span>

    <span class="c1">// For roots [NumMeta, NumRoots), the metadata pointer is null.</span>
    <span class="k">for</span> <span class="p">(</span><span class="kt">unsigned</span> <span class="n">e</span> <span class="o">=</span> <span class="n">R</span><span class="o">-&gt;</span><span class="n">Map</span><span class="o">-&gt;</span><span class="n">NumRoots</span><span class="p">;</span> <span class="n">i</span> <span class="o">!=</span> <span class="n">e</span><span class="p">;</span> <span class="o">++</span><span class="n">i</span><span class="p">)</span>
      <span class="n">Visitor</span><span class="p">(</span><span class="o">&amp;</span><span class="n">R</span><span class="o">-&gt;</span><span class="n">Roots</span><span class="p">[</span><span class="n">i</span><span class="p">],</span> <span class="nb">NULL</span><span class="p">);</span>
  <span class="p">}</span>
<span class="p">}</span>
</pre></div>
</div>
</div>
<div class="section" id="about-the-shadow-stack">
<h3><a class="toc-backref" href="#id6">About the shadow stack</a><a class="headerlink" href="#about-the-shadow-stack" title="Permalink to this headline"></a></h3>
<p>Unlike many GC algorithms which rely on a cooperative code generator to compile
stack maps, this algorithm carefully maintains a linked list of stack roots
[<a class="reference internal" href="#henderson02"><em>Henderson2002</em></a>].  This so-called &#8220;shadow stack&#8221; mirrors the
machine stack.  Maintaining this data structure is slower than using a stack map
compiled into the executable as constant data, but has a significant portability
advantage because it requires no special support from the target code generator,
and does not require tricky platform-specific code to crawl the machine stack.</p>
<p>The tradeoff for this simplicity and portability is:</p>
<ul class="simple">
<li>High overhead per function call.</li>
<li>Not thread-safe.</li>
</ul>
<p>Still, it&#8217;s an easy way to get started.  After your compiler and runtime are up
and running, writing a <a class="reference internal" href="#plugin"><em>plugin</em></a> will allow you to take advantage
of <a class="reference internal" href="#collector-algos"><em>more advanced GC features</em></a> of LLVM in order to
improve performance.</p>
</div>
</div>
<div class="section" id="ir-features">
<span id="gc-intrinsics"></span><h2><a class="toc-backref" href="#id7">IR features</a><a class="headerlink" href="#ir-features" title="Permalink to this headline"></a></h2>
<p>This section describes the garbage collection facilities provided by the
<a class="reference internal" href="LangRef.html"><em>LLVM intermediate representation</em></a>.  The exact behavior of these
IR features is specified by the binary interface implemented by a <a class="reference internal" href="#plugin"><em>code
generation plugin</em></a>, not by this document.</p>
<p>These facilities are limited to those strictly necessary; they are not intended
to be a complete interface to any garbage collector.  A program will need to
interface with the GC library using the facilities provided by that program.</p>
<div class="section" id="specifying-gc-code-generation-gc">
<h3><a class="toc-backref" href="#id8">Specifying GC code generation: <tt class="docutils literal"><span class="pre">gc</span> <span class="pre">&quot;...&quot;</span></tt></a><a class="headerlink" href="#specifying-gc-code-generation-gc" title="Permalink to this headline"></a></h3>
<div class="highlight-llvm"><div class="highlight"><pre>define ty @name(...) gc &quot;name&quot; { ...
</pre></div>
</div>
<p>The <tt class="docutils literal"><span class="pre">gc</span></tt> function attribute is used to specify the desired GC style to the
compiler.  Its programmatic equivalent is the <tt class="docutils literal"><span class="pre">setGC</span></tt> method of <tt class="docutils literal"><span class="pre">Function</span></tt>.</p>
<p>Setting <tt class="docutils literal"><span class="pre">gc</span> <span class="pre">&quot;name&quot;</span></tt> on a function triggers a search for a matching code
generation plugin &#8220;<em>name</em>&#8221;; it is that plugin which defines the exact nature of
the code generated to support GC.  If none is found, the compiler will raise an
error.</p>
<p>Specifying the GC style on a per-function basis allows LLVM to link together
programs that use different garbage collection algorithms (or none at all).</p>
</div>
<div class="section" id="gcroot">
<span id="identifying-gc-roots-on-the-stack-llvm-gcroot"></span><h3><a class="toc-backref" href="#id9">Identifying GC roots on the stack: <tt class="docutils literal"><span class="pre">llvm.gcroot</span></tt></a><a class="headerlink" href="#gcroot" title="Permalink to this headline"></a></h3>
<div class="highlight-llvm"><div class="highlight"><pre><span class="kt">void</span> <span class="vg">@llvm.gcroot</span><span class="p">(</span><span class="k">i8</span><span class="p">**</span> <span class="nv">%ptrloc</span><span class="p">,</span> <span class="k">i8</span><span class="p">*</span> <span class="nv">%metadata</span><span class="p">)</span>
</pre></div>
</div>
<p>The <tt class="docutils literal"><span class="pre">llvm.gcroot</span></tt> intrinsic is used to inform LLVM that a stack variable
references an object on the heap and is to be tracked for garbage collection.
The exact impact on generated code is specified by a <a class="reference internal" href="#plugin"><em>compiler plugin</em></a>.  All calls to <tt class="docutils literal"><span class="pre">llvm.gcroot</span></tt> <strong>must</strong> reside inside the first basic
block.</p>
<p>A compiler which uses mem2reg to raise imperative code using <tt class="docutils literal"><span class="pre">alloca</span></tt> into SSA
form need only add a call to <tt class="docutils literal"><span class="pre">&#64;llvm.gcroot</span></tt> for those variables which a
pointers into the GC heap.</p>
<p>It is also important to mark intermediate values with <tt class="docutils literal"><span class="pre">llvm.gcroot</span></tt>.  For
example, consider <tt class="docutils literal"><span class="pre">h(f(),</span> <span class="pre">g())</span></tt>.  Beware leaking the result of <tt class="docutils literal"><span class="pre">f()</span></tt> in the
case that <tt class="docutils literal"><span class="pre">g()</span></tt> triggers a collection.  Note, that stack variables must be
initialized and marked with <tt class="docutils literal"><span class="pre">llvm.gcroot</span></tt> in function&#8217;s prologue.</p>
<p>The first argument <strong>must</strong> be a value referring to an alloca instruction or a
bitcast of an alloca.  The second contains a pointer to metadata that should be
associated with the pointer, and <strong>must</strong> be a constant or global value
address.  If your target collector uses tags, use a null pointer for metadata.</p>
<p>The <tt class="docutils literal"><span class="pre">%metadata</span></tt> argument can be used to avoid requiring heap objects to have
&#8216;isa&#8217; pointers or tag bits. [<a class="reference internal" href="#appel89">Appel89</a>, <a class="reference internal" href="#goldberg91">Goldberg91</a>, <a class="reference internal" href="#tolmach94">Tolmach94</a>] If specified,
its value will be tracked along with the location of the pointer in the stack
frame.</p>
<p>Consider the following fragment of Java code:</p>
<div class="highlight-java"><div class="highlight"><pre><span class="o">{</span>
  <span class="n">Object</span> <span class="n">X</span><span class="o">;</span>   <span class="c1">// A null-initialized reference to an object</span>
  <span class="o">...</span>
<span class="o">}</span>
</pre></div>
</div>
<p>This block (which may be located in the middle of a function or in a loop nest),
could be compiled to this LLVM code:</p>
<div class="highlight-llvm"><div class="highlight"><pre><span class="nl">Entry:</span>
   <span class="c">;; In the entry block for the function, allocate the</span>
   <span class="c">;; stack space for X, which is an LLVM pointer.</span>
   <span class="nv">%X</span> <span class="p">=</span> <span class="k">alloca</span> <span class="nv">%Object</span><span class="p">*</span>

   <span class="c">;; Tell LLVM that the stack space is a stack root.</span>
   <span class="c">;; Java has type-tags on objects, so we pass null as metadata.</span>
   <span class="nv">%tmp</span> <span class="p">=</span> <span class="k">bitcast</span> <span class="nv">%Object</span><span class="p">**</span> <span class="nv">%X</span> <span class="k">to</span> <span class="k">i8</span><span class="p">**</span>
   <span class="k">call</span> <span class="kt">void</span> <span class="vg">@llvm.gcroot</span><span class="p">(</span><span class="k">i8</span><span class="p">**</span> <span class="nv">%tmp</span><span class="p">,</span> <span class="k">i8</span><span class="p">*</span> <span class="k">null</span><span class="p">)</span>
   <span class="p">...</span>

   <span class="c">;; &quot;CodeBlock&quot; is the block corresponding to the start</span>
   <span class="c">;;  of the scope above.</span>
<span class="nl">CodeBlock:</span>
   <span class="c">;; Java null-initializes pointers.</span>
   <span class="k">store</span> <span class="nv">%Object</span><span class="p">*</span> <span class="k">null</span><span class="p">,</span> <span class="nv">%Object</span><span class="p">**</span> <span class="nv">%X</span>

   <span class="p">...</span>

   <span class="c">;; As the pointer goes out of scope, store a null value into</span>
   <span class="c">;; it, to indicate that the value is no longer live.</span>
   <span class="k">store</span> <span class="nv">%Object</span><span class="p">*</span> <span class="k">null</span><span class="p">,</span> <span class="nv">%Object</span><span class="p">**</span> <span class="nv">%X</span>
   <span class="p">...</span>
</pre></div>
</div>
</div>
<div class="section" id="reading-and-writing-references-in-the-heap">
<h3><a class="toc-backref" href="#id10">Reading and writing references in the heap</a><a class="headerlink" href="#reading-and-writing-references-in-the-heap" title="Permalink to this headline"></a></h3>
<p>Some collectors need to be informed when the mutator (the program that needs
garbage collection) either reads a pointer from or writes a pointer to a field
of a heap object.  The code fragments inserted at these points are called <em>read
barriers</em> and <em>write barriers</em>, respectively.  The amount of code that needs to
be executed is usually quite small and not on the critical path of any
computation, so the overall performance impact of the barrier is tolerable.</p>
<p>Barriers often require access to the <em>object pointer</em> rather than the <em>derived
pointer</em> (which is a pointer to the field within the object).  Accordingly,
these intrinsics take both pointers as separate arguments for completeness.  In
this snippet, <tt class="docutils literal"><span class="pre">%object</span></tt> is the object pointer, and <tt class="docutils literal"><span class="pre">%derived</span></tt> is the derived
pointer:</p>
<div class="highlight-llvm"><div class="highlight"><pre><span class="c">;; An array type.</span>
<span class="nv">%class.Array</span> <span class="p">=</span> <span class="k">type</span> <span class="p">{</span> <span class="nv">%class.Object</span><span class="p">,</span> <span class="k">i32</span><span class="p">,</span> <span class="p">[</span><span class="m">0</span> <span class="k">x</span> <span class="nv">%class.Object</span><span class="p">*]</span> <span class="p">}</span>
<span class="p">...</span>

<span class="c">;; Load the object pointer from a gcroot.</span>
<span class="nv">%object</span> <span class="p">=</span> <span class="k">load</span> <span class="nv">%class.Array</span><span class="p">**</span> <span class="nv">%object_addr</span>

<span class="c">;; Compute the derived pointer.</span>
<span class="nv">%derived</span> <span class="p">=</span> <span class="k">getelementptr</span> <span class="nv">%object</span><span class="p">,</span> <span class="k">i32</span> <span class="m">0</span><span class="p">,</span> <span class="k">i32</span> <span class="m">2</span><span class="p">,</span> <span class="k">i32</span> <span class="nv">%n</span>
</pre></div>
</div>
<p>LLVM does not enforce this relationship between the object and derived pointer
(although a <a class="reference internal" href="#plugin"><em>plugin</em></a> might).  However, it would be an unusual
collector that violated it.</p>
<p>The use of these intrinsics is naturally optional if the target GC does require
the corresponding barrier.  Such a GC plugin will replace the intrinsic calls
with the corresponding <tt class="docutils literal"><span class="pre">load</span></tt> or <tt class="docutils literal"><span class="pre">store</span></tt> instruction if they are used.</p>
<div class="section" id="write-barrier-llvm-gcwrite">
<h4><a class="toc-backref" href="#id11">Write barrier: <tt class="docutils literal"><span class="pre">llvm.gcwrite</span></tt></a><a class="headerlink" href="#write-barrier-llvm-gcwrite" title="Permalink to this headline"></a></h4>
<div class="highlight-llvm"><div class="highlight"><pre><span class="kt">void</span> <span class="vg">@llvm.gcwrite</span><span class="p">(</span><span class="k">i8</span><span class="p">*</span> <span class="nv">%value</span><span class="p">,</span> <span class="k">i8</span><span class="p">*</span> <span class="nv">%object</span><span class="p">,</span> <span class="k">i8</span><span class="p">**</span> <span class="nv">%derived</span><span class="p">)</span>
</pre></div>
</div>
<p>For write barriers, LLVM provides the <tt class="docutils literal"><span class="pre">llvm.gcwrite</span></tt> intrinsic function.  It
has exactly the same semantics as a non-volatile <tt class="docutils literal"><span class="pre">store</span></tt> to the derived
pointer (the third argument).  The exact code generated is specified by a
compiler <a class="reference internal" href="#plugin"><em>plugin</em></a>.</p>
<p>Many important algorithms require write barriers, including generational and
concurrent collectors.  Additionally, write barriers could be used to implement
reference counting.</p>
</div>
<div class="section" id="read-barrier-llvm-gcread">
<h4><a class="toc-backref" href="#id12">Read barrier: <tt class="docutils literal"><span class="pre">llvm.gcread</span></tt></a><a class="headerlink" href="#read-barrier-llvm-gcread" title="Permalink to this headline"></a></h4>
<div class="highlight-llvm"><div class="highlight"><pre><span class="k">i8</span><span class="p">*</span> <span class="vg">@llvm.gcread</span><span class="p">(</span><span class="k">i8</span><span class="p">*</span> <span class="nv">%object</span><span class="p">,</span> <span class="k">i8</span><span class="p">**</span> <span class="nv">%derived</span><span class="p">)</span>
</pre></div>
</div>
<p>For read barriers, LLVM provides the <tt class="docutils literal"><span class="pre">llvm.gcread</span></tt> intrinsic function.  It has
exactly the same semantics as a non-volatile <tt class="docutils literal"><span class="pre">load</span></tt> from the derived pointer
(the second argument).  The exact code generated is specified by a
<a class="reference internal" href="#plugin"><em>compiler plugin</em></a>.</p>
<p>Read barriers are needed by fewer algorithms than write barriers, and may have a
greater performance impact since pointer reads are more frequent than writes.</p>
</div>
</div>
</div>
<div class="section" id="implementing-a-collector-plugin">
<span id="plugin"></span><h2><a class="toc-backref" href="#id13">Implementing a collector plugin</a><a class="headerlink" href="#implementing-a-collector-plugin" title="Permalink to this headline"></a></h2>
<p>User code specifies which GC code generation to use with the <tt class="docutils literal"><span class="pre">gc</span></tt> function
attribute or, equivalently, with the <tt class="docutils literal"><span class="pre">setGC</span></tt> method of <tt class="docutils literal"><span class="pre">Function</span></tt>.</p>
<p>To implement a GC plugin, it is necessary to subclass <tt class="docutils literal"><span class="pre">llvm::GCStrategy</span></tt>,
which can be accomplished in a few lines of boilerplate code.  LLVM&#8217;s
infrastructure provides access to several important algorithms.  For an
uncontroversial collector, all that remains may be to compile LLVM&#8217;s computed
stack map to assembly code (using the binary representation expected by the
runtime library).  This can be accomplished in about 100 lines of code.</p>
<p>This is not the appropriate place to implement a garbage collected heap or a
garbage collector itself.  That code should exist in the language&#8217;s runtime
library.  The compiler plugin is responsible for generating code which conforms
to the binary interface defined by library, most essentially the <a class="reference internal" href="#stack-map"><em>stack map</em></a>.</p>
<p>To subclass <tt class="docutils literal"><span class="pre">llvm::GCStrategy</span></tt> and register it with the compiler:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// lib/MyGC/MyGC.cpp - Example LLVM GC plugin</span>

<span class="cp">#include &quot;llvm/CodeGen/GCStrategy.h&quot;</span>
<span class="cp">#include &quot;llvm/CodeGen/GCMetadata.h&quot;</span>
<span class="cp">#include &quot;llvm/Support/Compiler.h&quot;</span>

<span class="k">using</span> <span class="k">namespace</span> <span class="n">llvm</span><span class="p">;</span>

<span class="k">namespace</span> <span class="p">{</span>
  <span class="k">class</span> <span class="nc">LLVM_LIBRARY_VISIBILITY</span> <span class="n">MyGC</span> <span class="o">:</span> <span class="k">public</span> <span class="n">GCStrategy</span> <span class="p">{</span>
  <span class="nl">public:</span>
    <span class="n">MyGC</span><span class="p">()</span> <span class="p">{}</span>
  <span class="p">};</span>

  <span class="n">GCRegistry</span><span class="o">::</span><span class="n">Add</span><span class="o">&lt;</span><span class="n">MyGC</span><span class="o">&gt;</span>
  <span class="n">X</span><span class="p">(</span><span class="s">&quot;mygc&quot;</span><span class="p">,</span> <span class="s">&quot;My bespoke garbage collector.&quot;</span><span class="p">);</span>
<span class="p">}</span>
</pre></div>
</div>
<p>This boilerplate collector does nothing.  More specifically:</p>
<ul class="simple">
<li><tt class="docutils literal"><span class="pre">llvm.gcread</span></tt> calls are replaced with the corresponding <tt class="docutils literal"><span class="pre">load</span></tt>
instruction.</li>
<li><tt class="docutils literal"><span class="pre">llvm.gcwrite</span></tt> calls are replaced with the corresponding <tt class="docutils literal"><span class="pre">store</span></tt>
instruction.</li>
<li>No safe points are added to the code.</li>
<li>The stack map is not compiled into the executable.</li>
</ul>
<p>Using the LLVM makefiles (like the <a class="reference external" href="http://llvm.org/viewvc/llvm-project/llvm/trunk/projects/sample/">sample project</a>), this code
can be compiled as a plugin using a simple makefile:</p>
<div class="highlight-make"><div class="highlight"><pre><span class="c"># lib/MyGC/Makefile</span>

<span class="nv">LEVEL</span> <span class="o">:=</span> ../..
<span class="nv">LIBRARYNAME</span> <span class="o">=</span> MyGC
<span class="nv">LOADABLE_MODULE</span> <span class="o">=</span> 1

<span class="cp">include $(LEVEL)/Makefile.common</span>
</pre></div>
</div>
<p>Once the plugin is compiled, code using it may be compiled using <tt class="docutils literal"><span class="pre">llc</span>
<span class="pre">-load=MyGC.so</span></tt> (though MyGC.so may have some other platform-specific
extension):</p>
<div class="highlight-python"><div class="highlight"><pre>$ cat sample.ll
define void @f() gc &quot;mygc&quot; {
entry:
  ret void
}
$ llvm-as &lt; sample.ll | llc -load=MyGC.so
</pre></div>
</div>
<p>It is also possible to statically link the collector plugin into tools, such as
a language-specific compiler front-end.</p>
<div class="section" id="overview-of-available-features">
<span id="collector-algos"></span><h3><a class="toc-backref" href="#id14">Overview of available features</a><a class="headerlink" href="#overview-of-available-features" title="Permalink to this headline"></a></h3>
<p><tt class="docutils literal"><span class="pre">GCStrategy</span></tt> provides a range of features through which a plugin may do useful
work.  Some of these are callbacks, some are algorithms that can be enabled,
disabled, or customized.  This matrix summarizes the supported (and planned)
features and correlates them with the collection techniques which typically
require them.</p>
<table border="1" class="docutils">
<colgroup>
<col width="14%" />
<col width="7%" />
<col width="9%" />
<col width="11%" />
<col width="8%" />
<col width="10%" />
<col width="15%" />
<col width="11%" />
<col width="14%" />
</colgroup>
<thead valign="bottom">
<tr class="row-odd"><th class="head">Algorithm</th>
<th class="head">Done</th>
<th class="head">Shadow
stack</th>
<th class="head">refcount</th>
<th class="head">mark-
sweep</th>
<th class="head">copying</th>
<th class="head">incremental</th>
<th class="head">threaded</th>
<th class="head">concurrent</th>
</tr>
</thead>
<tbody valign="top">
<tr class="row-even"><td>stack map</td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td></td>
<td></td>
<td></td>
<td></td>
<td></td>
</tr>
<tr class="row-odd"><td>initialize
roots</td>
<td></td>
<td></td>
<td></td>
<td></td>
<td></td>
<td></td>
<td></td>
<td></td>
</tr>
<tr class="row-even"><td>derived
pointers</td>
<td>NO</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td><strong>N</strong>*</td>
<td><strong>N</strong>*</td>
</tr>
<tr class="row-odd"><td><strong>custom
lowering</strong></td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
</tr>
<tr class="row-even"><td><em>gcroot</em></td>
<td></td>
<td></td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
</tr>
<tr class="row-odd"><td><em>gcwrite</em></td>
<td></td>
<td>&nbsp;</td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td></td>
<td>&nbsp;</td>
<td></td>
</tr>
<tr class="row-even"><td><em>gcread</em></td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td></td>
</tr>
<tr class="row-odd"><td><strong>safe
points</strong></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
</tr>
<tr class="row-even"><td><em>in
calls</em></td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td></td>
<td></td>
<td></td>
<td></td>
<td></td>
</tr>
<tr class="row-odd"><td><em>before
calls</em></td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td></td>
<td></td>
</tr>
<tr class="row-even"><td><em>for
loops</em></td>
<td>NO</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td><strong>N</strong></td>
<td><strong>N</strong></td>
</tr>
<tr class="row-odd"><td><em>before
escape</em></td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td></td>
<td></td>
</tr>
<tr class="row-even"><td>emit code
at safe
points</td>
<td>NO</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td><strong>N</strong></td>
<td><strong>N</strong></td>
</tr>
<tr class="row-odd"><td><strong>output</strong></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
</tr>
<tr class="row-even"><td><em>assembly</em></td>
<td></td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td></td>
<td></td>
<td></td>
<td></td>
<td></td>
</tr>
<tr class="row-odd"><td><em>JIT</em></td>
<td>NO</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
</tr>
<tr class="row-even"><td><em>obj</em></td>
<td>NO</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
</tr>
<tr class="row-odd"><td>live
analysis</td>
<td>NO</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
</tr>
<tr class="row-even"><td>register
map</td>
<td>NO</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
<td><strong>?</strong></td>
</tr>
<tr class="row-odd"><td colspan="9">* Derived pointers only pose a hasard to copying collections.</td>
</tr>
<tr class="row-even"><td colspan="9"><strong>?</strong> denotes a feature which could be utilized if available.</td>
</tr>
</tbody>
</table>
<p>To be clear, the collection techniques above are defined as:</p>
<dl class="docutils">
<dt>Shadow Stack</dt>
<dd>The mutator carefully maintains a linked list of stack roots.</dd>
<dt>Reference Counting</dt>
<dd>The mutator maintains a reference count for each object and frees an object
when its count falls to zero.</dd>
<dt>Mark-Sweep</dt>
<dd>When the heap is exhausted, the collector marks reachable objects starting
from the roots, then deallocates unreachable objects in a sweep phase.</dd>
<dt>Copying</dt>
<dd>As reachability analysis proceeds, the collector copies objects from one heap
area to another, compacting them in the process.  Copying collectors enable
highly efficient &#8220;bump pointer&#8221; allocation and can improve locality of
reference.</dd>
<dt>Incremental</dt>
<dd>(Including generational collectors.) Incremental collectors generally have all
the properties of a copying collector (regardless of whether the mature heap
is compacting), but bring the added complexity of requiring write barriers.</dd>
<dt>Threaded</dt>
<dd>Denotes a multithreaded mutator; the collector must still stop the mutator
(&#8220;stop the world&#8221;) before beginning reachability analysis.  Stopping a
multithreaded mutator is a complicated problem.  It generally requires highly
platform specific code in the runtime, and the production of carefully
designed machine code at safe points.</dd>
<dt>Concurrent</dt>
<dd>In this technique, the mutator and the collector run concurrently, with the
goal of eliminating pause times.  In a <em>cooperative</em> collector, the mutator
further aids with collection should a pause occur, allowing collection to take
advantage of multiprocessor hosts.  The &#8220;stop the world&#8221; problem of threaded
collectors is generally still present to a limited extent.  Sophisticated
marking algorithms are necessary.  Read barriers may be necessary.</dd>
</dl>
<p>As the matrix indicates, LLVM&#8217;s garbage collection infrastructure is already
suitable for a wide variety of collectors, but does not currently extend to
multithreaded programs.  This will be added in the future as there is
interest.</p>
</div>
<div class="section" id="computing-stack-maps">
<span id="stack-map"></span><h3><a class="toc-backref" href="#id15">Computing stack maps</a><a class="headerlink" href="#computing-stack-maps" title="Permalink to this headline"></a></h3>
<p>LLVM automatically computes a stack map.  One of the most important features
of a <tt class="docutils literal"><span class="pre">GCStrategy</span></tt> is to compile this information into the executable in
the binary representation expected by the runtime library.</p>
<p>The stack map consists of the location and identity of each GC root in the
each function in the module.  For each root:</p>
<ul class="simple">
<li><tt class="docutils literal"><span class="pre">RootNum</span></tt>: The index of the root.</li>
<li><tt class="docutils literal"><span class="pre">StackOffset</span></tt>: The offset of the object relative to the frame pointer.</li>
<li><tt class="docutils literal"><span class="pre">RootMetadata</span></tt>: The value passed as the <tt class="docutils literal"><span class="pre">%metadata</span></tt> parameter to the
<tt class="docutils literal"><span class="pre">&#64;llvm.gcroot</span></tt> intrinsic.</li>
</ul>
<p>Also, for the function as a whole:</p>
<ul>
<li><dl class="first docutils">
<dt><tt class="docutils literal"><span class="pre">getFrameSize()</span></tt>: The overall size of the function&#8217;s initial stack frame,</dt>
<dd><p class="first last">not accounting for any dynamic allocation.</p>
</dd>
</dl>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">roots_size()</span></tt>: The count of roots in the function.</p>
</li>
</ul>
<p>To access the stack map, use <tt class="docutils literal"><span class="pre">GCFunctionMetadata::roots_begin()</span></tt> and
-<tt class="docutils literal"><span class="pre">end()</span></tt> from the <a class="reference internal" href="#assembly"><em>GCMetadataPrinter</em></a>:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="k">for</span> <span class="p">(</span><span class="n">iterator</span> <span class="n">I</span> <span class="o">=</span> <span class="n">begin</span><span class="p">(),</span> <span class="n">E</span> <span class="o">=</span> <span class="n">end</span><span class="p">();</span> <span class="n">I</span> <span class="o">!=</span> <span class="n">E</span><span class="p">;</span> <span class="o">++</span><span class="n">I</span><span class="p">)</span> <span class="p">{</span>
  <span class="n">GCFunctionInfo</span> <span class="o">*</span><span class="n">FI</span> <span class="o">=</span> <span class="o">*</span><span class="n">I</span><span class="p">;</span>
  <span class="kt">unsigned</span> <span class="n">FrameSize</span> <span class="o">=</span> <span class="n">FI</span><span class="o">-&gt;</span><span class="n">getFrameSize</span><span class="p">();</span>
  <span class="kt">size_t</span> <span class="n">RootCount</span> <span class="o">=</span> <span class="n">FI</span><span class="o">-&gt;</span><span class="n">roots_size</span><span class="p">();</span>

  <span class="k">for</span> <span class="p">(</span><span class="n">GCFunctionInfo</span><span class="o">::</span><span class="n">roots_iterator</span> <span class="n">RI</span> <span class="o">=</span> <span class="n">FI</span><span class="o">-&gt;</span><span class="n">roots_begin</span><span class="p">(),</span>
                                      <span class="n">RE</span> <span class="o">=</span> <span class="n">FI</span><span class="o">-&gt;</span><span class="n">roots_end</span><span class="p">();</span>
                                      <span class="n">RI</span> <span class="o">!=</span> <span class="n">RE</span><span class="p">;</span> <span class="o">++</span><span class="n">RI</span><span class="p">)</span> <span class="p">{</span>
    <span class="kt">int</span> <span class="n">RootNum</span> <span class="o">=</span> <span class="n">RI</span><span class="o">-&gt;</span><span class="n">Num</span><span class="p">;</span>
    <span class="kt">int</span> <span class="n">RootStackOffset</span> <span class="o">=</span> <span class="n">RI</span><span class="o">-&gt;</span><span class="n">StackOffset</span><span class="p">;</span>
    <span class="n">Constant</span> <span class="o">*</span><span class="n">RootMetadata</span> <span class="o">=</span> <span class="n">RI</span><span class="o">-&gt;</span><span class="n">Metadata</span><span class="p">;</span>
  <span class="p">}</span>
<span class="p">}</span>
</pre></div>
</div>
<p>If the <tt class="docutils literal"><span class="pre">llvm.gcroot</span></tt> intrinsic is eliminated before code generation by a
custom lowering pass, LLVM will compute an empty stack map.  This may be useful
for collector plugins which implement reference counting or a shadow stack.</p>
</div>
<div class="section" id="initializing-roots-to-null-initroots">
<span id="init-roots"></span><h3><a class="toc-backref" href="#id16">Initializing roots to null: <tt class="docutils literal"><span class="pre">InitRoots</span></tt></a><a class="headerlink" href="#initializing-roots-to-null-initroots" title="Permalink to this headline"></a></h3>
<div class="highlight-c++"><div class="highlight"><pre><span class="n">MyGC</span><span class="o">::</span><span class="n">MyGC</span><span class="p">()</span> <span class="p">{</span>
  <span class="n">InitRoots</span> <span class="o">=</span> <span class="nb">true</span><span class="p">;</span>
<span class="p">}</span>
</pre></div>
</div>
<p>When set, LLVM will automatically initialize each root to <tt class="docutils literal"><span class="pre">null</span></tt> upon entry to
the function.  This prevents the GC&#8217;s sweep phase from visiting uninitialized
pointers, which will almost certainly cause it to crash.  This initialization
occurs before custom lowering, so the two may be used together.</p>
<p>Since LLVM does not yet compute liveness information, there is no means of
distinguishing an uninitialized stack root from an initialized one.  Therefore,
this feature should be used by all GC plugins.  It is enabled by default.</p>
</div>
<div class="section" id="custom-lowering-of-intrinsics-customroots-customreadbarriers-and-customwritebarriers">
<h3><a class="toc-backref" href="#id17">Custom lowering of intrinsics: <tt class="docutils literal"><span class="pre">CustomRoots</span></tt>, <tt class="docutils literal"><span class="pre">CustomReadBarriers</span></tt>, and <tt class="docutils literal"><span class="pre">CustomWriteBarriers</span></tt></a><a class="headerlink" href="#custom-lowering-of-intrinsics-customroots-customreadbarriers-and-customwritebarriers" title="Permalink to this headline"></a></h3>
<p>For GCs which use barriers or unusual treatment of stack roots, these flags
allow the collector to perform arbitrary transformations of the LLVM IR:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="k">class</span> <span class="nc">MyGC</span> <span class="o">:</span> <span class="k">public</span> <span class="n">GCStrategy</span> <span class="p">{</span>
<span class="nl">public:</span>
  <span class="n">MyGC</span><span class="p">()</span> <span class="p">{</span>
    <span class="n">CustomRoots</span> <span class="o">=</span> <span class="nb">true</span><span class="p">;</span>
    <span class="n">CustomReadBarriers</span> <span class="o">=</span> <span class="nb">true</span><span class="p">;</span>
    <span class="n">CustomWriteBarriers</span> <span class="o">=</span> <span class="nb">true</span><span class="p">;</span>
  <span class="p">}</span>

  <span class="k">virtual</span> <span class="kt">bool</span> <span class="n">initializeCustomLowering</span><span class="p">(</span><span class="n">Module</span> <span class="o">&amp;</span><span class="n">M</span><span class="p">);</span>
  <span class="k">virtual</span> <span class="kt">bool</span> <span class="nf">performCustomLowering</span><span class="p">(</span><span class="n">Function</span> <span class="o">&amp;</span><span class="n">F</span><span class="p">);</span>
<span class="p">};</span>
</pre></div>
</div>
<p>If any of these flags are set, then LLVM suppresses its default lowering for the
corresponding intrinsics and instead calls <tt class="docutils literal"><span class="pre">performCustomLowering</span></tt>.</p>
<p>LLVM&#8217;s default action for each intrinsic is as follows:</p>
<ul class="simple">
<li><tt class="docutils literal"><span class="pre">llvm.gcroot</span></tt>: Leave it alone.  The code generator must see it or the stack
map will not be computed.</li>
<li><tt class="docutils literal"><span class="pre">llvm.gcread</span></tt>: Substitute a <tt class="docutils literal"><span class="pre">load</span></tt> instruction.</li>
<li><tt class="docutils literal"><span class="pre">llvm.gcwrite</span></tt>: Substitute a <tt class="docutils literal"><span class="pre">store</span></tt> instruction.</li>
</ul>
<p>If <tt class="docutils literal"><span class="pre">CustomReadBarriers</span></tt> or <tt class="docutils literal"><span class="pre">CustomWriteBarriers</span></tt> are specified, then
<tt class="docutils literal"><span class="pre">performCustomLowering</span></tt> <strong>must</strong> eliminate the corresponding barriers.</p>
<p><tt class="docutils literal"><span class="pre">performCustomLowering</span></tt> must comply with the same restrictions as
<a class="reference internal" href="WritingAnLLVMPass.html#writing-an-llvm-pass-runonfunction"><em>FunctionPass::runOnFunction</em></a>
Likewise, <tt class="docutils literal"><span class="pre">initializeCustomLowering</span></tt> has the same semantics as
<a class="reference internal" href="WritingAnLLVMPass.html#writing-an-llvm-pass-doinitialization-mod"><em>Pass::doInitialization(Module&amp;)</em></a></p>
<p>The following can be used as a template:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include &quot;llvm/IR/Module.h&quot;</span>
<span class="cp">#include &quot;llvm/IR/IntrinsicInst.h&quot;</span>

<span class="kt">bool</span> <span class="n">MyGC</span><span class="o">::</span><span class="n">initializeCustomLowering</span><span class="p">(</span><span class="n">Module</span> <span class="o">&amp;</span><span class="n">M</span><span class="p">)</span> <span class="p">{</span>
  <span class="k">return</span> <span class="nb">false</span><span class="p">;</span>
<span class="p">}</span>

<span class="kt">bool</span> <span class="n">MyGC</span><span class="o">::</span><span class="n">performCustomLowering</span><span class="p">(</span><span class="n">Function</span> <span class="o">&amp;</span><span class="n">F</span><span class="p">)</span> <span class="p">{</span>
  <span class="kt">bool</span> <span class="n">MadeChange</span> <span class="o">=</span> <span class="nb">false</span><span class="p">;</span>

  <span class="k">for</span> <span class="p">(</span><span class="n">Function</span><span class="o">::</span><span class="n">iterator</span> <span class="n">BB</span> <span class="o">=</span> <span class="n">F</span><span class="p">.</span><span class="n">begin</span><span class="p">(),</span> <span class="n">E</span> <span class="o">=</span> <span class="n">F</span><span class="p">.</span><span class="n">end</span><span class="p">();</span> <span class="n">BB</span> <span class="o">!=</span> <span class="n">E</span><span class="p">;</span> <span class="o">++</span><span class="n">BB</span><span class="p">)</span>
    <span class="k">for</span> <span class="p">(</span><span class="n">BasicBlock</span><span class="o">::</span><span class="n">iterator</span> <span class="n">II</span> <span class="o">=</span> <span class="n">BB</span><span class="o">-&gt;</span><span class="n">begin</span><span class="p">(),</span> <span class="n">E</span> <span class="o">=</span> <span class="n">BB</span><span class="o">-&gt;</span><span class="n">end</span><span class="p">();</span> <span class="n">II</span> <span class="o">!=</span> <span class="n">E</span><span class="p">;</span> <span class="p">)</span>
      <span class="k">if</span> <span class="p">(</span><span class="n">IntrinsicInst</span> <span class="o">*</span><span class="n">CI</span> <span class="o">=</span> <span class="n">dyn_cast</span><span class="o">&lt;</span><span class="n">IntrinsicInst</span><span class="o">&gt;</span><span class="p">(</span><span class="n">II</span><span class="o">++</span><span class="p">))</span>
        <span class="k">if</span> <span class="p">(</span><span class="n">Function</span> <span class="o">*</span><span class="n">F</span> <span class="o">=</span> <span class="n">CI</span><span class="o">-&gt;</span><span class="n">getCalledFunction</span><span class="p">())</span>
          <span class="k">switch</span> <span class="p">(</span><span class="n">F</span><span class="o">-&gt;</span><span class="n">getIntrinsicID</span><span class="p">())</span> <span class="p">{</span>
          <span class="k">case</span> <span class="n">Intrinsic</span>:<span class="o">:</span><span class="n">gcwrite</span><span class="o">:</span>
            <span class="c1">// Handle llvm.gcwrite.</span>
            <span class="n">CI</span><span class="o">-&gt;</span><span class="n">eraseFromParent</span><span class="p">();</span>
            <span class="n">MadeChange</span> <span class="o">=</span> <span class="nb">true</span><span class="p">;</span>
            <span class="k">break</span><span class="p">;</span>
          <span class="k">case</span> <span class="n">Intrinsic</span>:<span class="o">:</span><span class="n">gcread</span><span class="o">:</span>
            <span class="c1">// Handle llvm.gcread.</span>
            <span class="n">CI</span><span class="o">-&gt;</span><span class="n">eraseFromParent</span><span class="p">();</span>
            <span class="n">MadeChange</span> <span class="o">=</span> <span class="nb">true</span><span class="p">;</span>
            <span class="k">break</span><span class="p">;</span>
          <span class="k">case</span> <span class="n">Intrinsic</span>:<span class="o">:</span><span class="n">gcroot</span><span class="o">:</span>
            <span class="c1">// Handle llvm.gcroot.</span>
            <span class="n">CI</span><span class="o">-&gt;</span><span class="n">eraseFromParent</span><span class="p">();</span>
            <span class="n">MadeChange</span> <span class="o">=</span> <span class="nb">true</span><span class="p">;</span>
            <span class="k">break</span><span class="p">;</span>
          <span class="p">}</span>

  <span class="k">return</span> <span class="n">MadeChange</span><span class="p">;</span>
<span class="p">}</span>
</pre></div>
</div>
</div>
<div class="section" id="generating-safe-points-neededsafepoints">
<span id="safe-points"></span><h3><a class="toc-backref" href="#id18">Generating safe points: <tt class="docutils literal"><span class="pre">NeededSafePoints</span></tt></a><a class="headerlink" href="#generating-safe-points-neededsafepoints" title="Permalink to this headline"></a></h3>
<p>LLVM can compute four kinds of safe points:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="k">namespace</span> <span class="n">GC</span> <span class="p">{</span>
  <span class="c1">/// PointKind - The type of a collector-safe point.</span>
  <span class="c1">///</span>
  <span class="k">enum</span> <span class="n">PointKind</span> <span class="p">{</span>
    <span class="n">Loop</span><span class="p">,</span>    <span class="c1">//&lt; Instr is a loop (backwards branch).</span>
    <span class="n">Return</span><span class="p">,</span>  <span class="c1">//&lt; Instr is a return instruction.</span>
    <span class="n">PreCall</span><span class="p">,</span> <span class="c1">//&lt; Instr is a call instruction.</span>
    <span class="n">PostCall</span> <span class="c1">//&lt; Instr is the return address of a call.</span>
  <span class="p">};</span>
<span class="p">}</span>
</pre></div>
</div>
<p>A collector can request any combination of the four by setting the
<tt class="docutils literal"><span class="pre">NeededSafePoints</span></tt> mask:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="n">MyGC</span><span class="o">::</span><span class="n">MyGC</span><span class="p">()</span>  <span class="p">{</span>
  <span class="n">NeededSafePoints</span> <span class="o">=</span> <span class="mi">1</span> <span class="o">&lt;&lt;</span> <span class="n">GC</span><span class="o">::</span><span class="n">Loop</span>
                   <span class="o">|</span> <span class="mi">1</span> <span class="o">&lt;&lt;</span> <span class="n">GC</span><span class="o">::</span><span class="n">Return</span>
                   <span class="o">|</span> <span class="mi">1</span> <span class="o">&lt;&lt;</span> <span class="n">GC</span><span class="o">::</span><span class="n">PreCall</span>
                   <span class="o">|</span> <span class="mi">1</span> <span class="o">&lt;&lt;</span> <span class="n">GC</span><span class="o">::</span><span class="n">PostCall</span><span class="p">;</span>
<span class="p">}</span>
</pre></div>
</div>
<p>It can then use the following routines to access safe points.</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="k">for</span> <span class="p">(</span><span class="n">iterator</span> <span class="n">I</span> <span class="o">=</span> <span class="n">begin</span><span class="p">(),</span> <span class="n">E</span> <span class="o">=</span> <span class="n">end</span><span class="p">();</span> <span class="n">I</span> <span class="o">!=</span> <span class="n">E</span><span class="p">;</span> <span class="o">++</span><span class="n">I</span><span class="p">)</span> <span class="p">{</span>
  <span class="n">GCFunctionInfo</span> <span class="o">*</span><span class="n">MD</span> <span class="o">=</span> <span class="o">*</span><span class="n">I</span><span class="p">;</span>
  <span class="kt">size_t</span> <span class="n">PointCount</span> <span class="o">=</span> <span class="n">MD</span><span class="o">-&gt;</span><span class="n">size</span><span class="p">();</span>

  <span class="k">for</span> <span class="p">(</span><span class="n">GCFunctionInfo</span><span class="o">::</span><span class="n">iterator</span> <span class="n">PI</span> <span class="o">=</span> <span class="n">MD</span><span class="o">-&gt;</span><span class="n">begin</span><span class="p">(),</span>
                                <span class="n">PE</span> <span class="o">=</span> <span class="n">MD</span><span class="o">-&gt;</span><span class="n">end</span><span class="p">();</span> <span class="n">PI</span> <span class="o">!=</span> <span class="n">PE</span><span class="p">;</span> <span class="o">++</span><span class="n">PI</span><span class="p">)</span> <span class="p">{</span>
    <span class="n">GC</span><span class="o">::</span><span class="n">PointKind</span> <span class="n">PointKind</span> <span class="o">=</span> <span class="n">PI</span><span class="o">-&gt;</span><span class="n">Kind</span><span class="p">;</span>
    <span class="kt">unsigned</span> <span class="n">PointNum</span> <span class="o">=</span> <span class="n">PI</span><span class="o">-&gt;</span><span class="n">Num</span><span class="p">;</span>
  <span class="p">}</span>
<span class="p">}</span>
</pre></div>
</div>
<p>Almost every collector requires <tt class="docutils literal"><span class="pre">PostCall</span></tt> safe points, since these correspond
to the moments when the function is suspended during a call to a subroutine.</p>
<p>Threaded programs generally require <tt class="docutils literal"><span class="pre">Loop</span></tt> safe points to guarantee that the
application will reach a safe point within a bounded amount of time, even if it
is executing a long-running loop which contains no function calls.</p>
<p>Threaded collectors may also require <tt class="docutils literal"><span class="pre">Return</span></tt> and <tt class="docutils literal"><span class="pre">PreCall</span></tt> safe points to
implement &#8220;stop the world&#8221; techniques using self-modifying code, where it is
important that the program not exit the function without reaching a safe point
(because only the topmost function has been patched).</p>
</div>
<div class="section" id="emitting-assembly-code-gcmetadataprinter">
<span id="assembly"></span><h3><a class="toc-backref" href="#id19">Emitting assembly code: <tt class="docutils literal"><span class="pre">GCMetadataPrinter</span></tt></a><a class="headerlink" href="#emitting-assembly-code-gcmetadataprinter" title="Permalink to this headline"></a></h3>
<p>LLVM allows a plugin to print arbitrary assembly code before and after the rest
of a module&#8217;s assembly code.  At the end of the module, the GC can compile the
LLVM stack map into assembly code. (At the beginning, this information is not
yet computed.)</p>
<p>Since AsmWriter and CodeGen are separate components of LLVM, a separate abstract
base class and registry is provided for printing assembly code, the
<tt class="docutils literal"><span class="pre">GCMetadaPrinter</span></tt> and <tt class="docutils literal"><span class="pre">GCMetadataPrinterRegistry</span></tt>.  The AsmWriter will look
for such a subclass if the <tt class="docutils literal"><span class="pre">GCStrategy</span></tt> sets <tt class="docutils literal"><span class="pre">UsesMetadata</span></tt>:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="n">MyGC</span><span class="o">::</span><span class="n">MyGC</span><span class="p">()</span> <span class="p">{</span>
  <span class="n">UsesMetadata</span> <span class="o">=</span> <span class="nb">true</span><span class="p">;</span>
<span class="p">}</span>
</pre></div>
</div>
<p>This separation allows JIT-only clients to be smaller.</p>
<p>Note that LLVM does not currently have analogous APIs to support code generation
in the JIT, nor using the object writers.</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="c1">// lib/MyGC/MyGCPrinter.cpp - Example LLVM GC printer</span>

<span class="cp">#include &quot;llvm/CodeGen/GCMetadataPrinter.h&quot;</span>
<span class="cp">#include &quot;llvm/Support/Compiler.h&quot;</span>

<span class="k">using</span> <span class="k">namespace</span> <span class="n">llvm</span><span class="p">;</span>

<span class="k">namespace</span> <span class="p">{</span>
  <span class="k">class</span> <span class="nc">LLVM_LIBRARY_VISIBILITY</span> <span class="n">MyGCPrinter</span> <span class="o">:</span> <span class="k">public</span> <span class="n">GCMetadataPrinter</span> <span class="p">{</span>
  <span class="nl">public:</span>
    <span class="k">virtual</span> <span class="kt">void</span> <span class="n">beginAssembly</span><span class="p">(</span><span class="n">AsmPrinter</span> <span class="o">&amp;</span><span class="n">AP</span><span class="p">);</span>

    <span class="k">virtual</span> <span class="kt">void</span> <span class="nf">finishAssembly</span><span class="p">(</span><span class="n">AsmPrinter</span> <span class="o">&amp;</span><span class="n">AP</span><span class="p">);</span>
  <span class="p">};</span>

  <span class="n">GCMetadataPrinterRegistry</span><span class="o">::</span><span class="n">Add</span><span class="o">&lt;</span><span class="n">MyGCPrinter</span><span class="o">&gt;</span>
  <span class="n">X</span><span class="p">(</span><span class="s">&quot;mygc&quot;</span><span class="p">,</span> <span class="s">&quot;My bespoke garbage collector.&quot;</span><span class="p">);</span>
<span class="p">}</span>
</pre></div>
</div>
<p>The collector should use <tt class="docutils literal"><span class="pre">AsmPrinter</span></tt> to print portable assembly code.  The
collector itself contains the stack map for the entire module, and may access
the <tt class="docutils literal"><span class="pre">GCFunctionInfo</span></tt> using its own <tt class="docutils literal"><span class="pre">begin()</span></tt> and <tt class="docutils literal"><span class="pre">end()</span></tt> methods.  Here&#8217;s
a realistic example:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="cp">#include &quot;llvm/CodeGen/AsmPrinter.h&quot;</span>
<span class="cp">#include &quot;llvm/IR/Function.h&quot;</span>
<span class="cp">#include &quot;llvm/IR/DataLayout.h&quot;</span>
<span class="cp">#include &quot;llvm/Target/TargetAsmInfo.h&quot;</span>
<span class="cp">#include &quot;llvm/Target/TargetMachine.h&quot;</span>

<span class="kt">void</span> <span class="n">MyGCPrinter</span><span class="o">::</span><span class="n">beginAssembly</span><span class="p">(</span><span class="n">AsmPrinter</span> <span class="o">&amp;</span><span class="n">AP</span><span class="p">)</span> <span class="p">{</span>
  <span class="c1">// Nothing to do.</span>
<span class="p">}</span>

<span class="kt">void</span> <span class="n">MyGCPrinter</span><span class="o">::</span><span class="n">finishAssembly</span><span class="p">(</span><span class="n">AsmPrinter</span> <span class="o">&amp;</span><span class="n">AP</span><span class="p">)</span> <span class="p">{</span>
  <span class="n">MCStreamer</span> <span class="o">&amp;</span><span class="n">OS</span> <span class="o">=</span> <span class="n">AP</span><span class="p">.</span><span class="n">OutStreamer</span><span class="p">;</span>
  <span class="kt">unsigned</span> <span class="n">IntPtrSize</span> <span class="o">=</span> <span class="n">AP</span><span class="p">.</span><span class="n">TM</span><span class="p">.</span><span class="n">getDataLayout</span><span class="p">()</span><span class="o">-&gt;</span><span class="n">getPointerSize</span><span class="p">();</span>

  <span class="c1">// Put this in the data section.</span>
  <span class="n">OS</span><span class="p">.</span><span class="n">SwitchSection</span><span class="p">(</span><span class="n">AP</span><span class="p">.</span><span class="n">getObjFileLowering</span><span class="p">().</span><span class="n">getDataSection</span><span class="p">());</span>

  <span class="c1">// For each function...</span>
  <span class="k">for</span> <span class="p">(</span><span class="n">iterator</span> <span class="n">FI</span> <span class="o">=</span> <span class="n">begin</span><span class="p">(),</span> <span class="n">FE</span> <span class="o">=</span> <span class="n">end</span><span class="p">();</span> <span class="n">FI</span> <span class="o">!=</span> <span class="n">FE</span><span class="p">;</span> <span class="o">++</span><span class="n">FI</span><span class="p">)</span> <span class="p">{</span>
    <span class="n">GCFunctionInfo</span> <span class="o">&amp;</span><span class="n">MD</span> <span class="o">=</span> <span class="o">**</span><span class="n">FI</span><span class="p">;</span>

    <span class="c1">// A compact GC layout. Emit this data structure:</span>
    <span class="c1">//</span>
    <span class="c1">// struct {</span>
    <span class="c1">//   int32_t PointCount;</span>
    <span class="c1">//   void *SafePointAddress[PointCount];</span>
    <span class="c1">//   int32_t StackFrameSize; // in words</span>
    <span class="c1">//   int32_t StackArity;</span>
    <span class="c1">//   int32_t LiveCount;</span>
    <span class="c1">//   int32_t LiveOffsets[LiveCount];</span>
    <span class="c1">// } __gcmap_&lt;FUNCTIONNAME&gt;;</span>

    <span class="c1">// Align to address width.</span>
    <span class="n">AP</span><span class="p">.</span><span class="n">EmitAlignment</span><span class="p">(</span><span class="n">IntPtrSize</span> <span class="o">==</span> <span class="mi">4</span> <span class="o">?</span> <span class="mi">2</span> <span class="o">:</span> <span class="mi">3</span><span class="p">);</span>

    <span class="c1">// Emit PointCount.</span>
    <span class="n">OS</span><span class="p">.</span><span class="n">AddComment</span><span class="p">(</span><span class="s">&quot;safe point count&quot;</span><span class="p">);</span>
    <span class="n">AP</span><span class="p">.</span><span class="n">EmitInt32</span><span class="p">(</span><span class="n">MD</span><span class="p">.</span><span class="n">size</span><span class="p">());</span>

    <span class="c1">// And each safe point...</span>
    <span class="k">for</span> <span class="p">(</span><span class="n">GCFunctionInfo</span><span class="o">::</span><span class="n">iterator</span> <span class="n">PI</span> <span class="o">=</span> <span class="n">MD</span><span class="p">.</span><span class="n">begin</span><span class="p">(),</span>
                                  <span class="n">PE</span> <span class="o">=</span> <span class="n">MD</span><span class="p">.</span><span class="n">end</span><span class="p">();</span> <span class="n">PI</span> <span class="o">!=</span> <span class="n">PE</span><span class="p">;</span> <span class="o">++</span><span class="n">PI</span><span class="p">)</span> <span class="p">{</span>
      <span class="c1">// Emit the address of the safe point.</span>
      <span class="n">OS</span><span class="p">.</span><span class="n">AddComment</span><span class="p">(</span><span class="s">&quot;safe point address&quot;</span><span class="p">);</span>
      <span class="n">MCSymbol</span> <span class="o">*</span><span class="n">Label</span> <span class="o">=</span> <span class="n">PI</span><span class="o">-&gt;</span><span class="n">Label</span><span class="p">;</span>
      <span class="n">AP</span><span class="p">.</span><span class="n">EmitLabelPlusOffset</span><span class="p">(</span><span class="n">Label</span><span class="cm">/*Hi*/</span><span class="p">,</span> <span class="mi">0</span><span class="cm">/*Offset*/</span><span class="p">,</span> <span class="mi">4</span><span class="cm">/*Size*/</span><span class="p">);</span>
    <span class="p">}</span>

    <span class="c1">// Stack information never change in safe points! Only print info from the</span>
    <span class="c1">// first call-site.</span>
    <span class="n">GCFunctionInfo</span><span class="o">::</span><span class="n">iterator</span> <span class="n">PI</span> <span class="o">=</span> <span class="n">MD</span><span class="p">.</span><span class="n">begin</span><span class="p">();</span>

    <span class="c1">// Emit the stack frame size.</span>
    <span class="n">OS</span><span class="p">.</span><span class="n">AddComment</span><span class="p">(</span><span class="s">&quot;stack frame size (in words)&quot;</span><span class="p">);</span>
    <span class="n">AP</span><span class="p">.</span><span class="n">EmitInt32</span><span class="p">(</span><span class="n">MD</span><span class="p">.</span><span class="n">getFrameSize</span><span class="p">()</span> <span class="o">/</span> <span class="n">IntPtrSize</span><span class="p">);</span>

    <span class="c1">// Emit stack arity, i.e. the number of stacked arguments.</span>
    <span class="kt">unsigned</span> <span class="n">RegisteredArgs</span> <span class="o">=</span> <span class="n">IntPtrSize</span> <span class="o">==</span> <span class="mi">4</span> <span class="o">?</span> <span class="mi">5</span> <span class="o">:</span> <span class="mi">6</span><span class="p">;</span>
    <span class="kt">unsigned</span> <span class="n">StackArity</span> <span class="o">=</span> <span class="n">MD</span><span class="p">.</span><span class="n">getFunction</span><span class="p">().</span><span class="n">arg_size</span><span class="p">()</span> <span class="o">&gt;</span> <span class="n">RegisteredArgs</span> <span class="o">?</span>
                          <span class="n">MD</span><span class="p">.</span><span class="n">getFunction</span><span class="p">().</span><span class="n">arg_size</span><span class="p">()</span> <span class="o">-</span> <span class="n">RegisteredArgs</span> <span class="o">:</span> <span class="mi">0</span><span class="p">;</span>
    <span class="n">OS</span><span class="p">.</span><span class="n">AddComment</span><span class="p">(</span><span class="s">&quot;stack arity&quot;</span><span class="p">);</span>
    <span class="n">AP</span><span class="p">.</span><span class="n">EmitInt32</span><span class="p">(</span><span class="n">StackArity</span><span class="p">);</span>

    <span class="c1">// Emit the number of live roots in the function.</span>
    <span class="n">OS</span><span class="p">.</span><span class="n">AddComment</span><span class="p">(</span><span class="s">&quot;live root count&quot;</span><span class="p">);</span>
    <span class="n">AP</span><span class="p">.</span><span class="n">EmitInt32</span><span class="p">(</span><span class="n">MD</span><span class="p">.</span><span class="n">live_size</span><span class="p">(</span><span class="n">PI</span><span class="p">));</span>

    <span class="c1">// And for each live root...</span>
    <span class="k">for</span> <span class="p">(</span><span class="n">GCFunctionInfo</span><span class="o">::</span><span class="n">live_iterator</span> <span class="n">LI</span> <span class="o">=</span> <span class="n">MD</span><span class="p">.</span><span class="n">live_begin</span><span class="p">(</span><span class="n">PI</span><span class="p">),</span>
                                       <span class="n">LE</span> <span class="o">=</span> <span class="n">MD</span><span class="p">.</span><span class="n">live_end</span><span class="p">(</span><span class="n">PI</span><span class="p">);</span>
                                       <span class="n">LI</span> <span class="o">!=</span> <span class="n">LE</span><span class="p">;</span> <span class="o">++</span><span class="n">LI</span><span class="p">)</span> <span class="p">{</span>
      <span class="c1">// Emit live root&#39;s offset within the stack frame.</span>
      <span class="n">OS</span><span class="p">.</span><span class="n">AddComment</span><span class="p">(</span><span class="s">&quot;stack index (offset / wordsize)&quot;</span><span class="p">);</span>
      <span class="n">AP</span><span class="p">.</span><span class="n">EmitInt32</span><span class="p">(</span><span class="n">LI</span><span class="o">-&gt;</span><span class="n">StackOffset</span><span class="p">);</span>
    <span class="p">}</span>
  <span class="p">}</span>
<span class="p">}</span>
</pre></div>
</div>
</div>
</div>
<div class="section" id="references">
<h2><a class="toc-backref" href="#id20">References</a><a class="headerlink" href="#references" title="Permalink to this headline"></a></h2>
<p id="appel89">[Appel89] Runtime Tags Aren&#8217;t Necessary. Andrew W. Appel. Lisp and Symbolic
Computation 19(7):703-705, July 1989.</p>
<p id="goldberg91">[Goldberg91] Tag-free garbage collection for strongly typed programming
languages. Benjamin Goldberg. ACM SIGPLAN PLDI&#8216;91.</p>
<p id="tolmach94">[Tolmach94] Tag-free garbage collection using explicit type parameters. Andrew
Tolmach. Proceedings of the 1994 ACM conference on LISP and functional
programming.</p>
<p id="henderson02">[Henderson2002] <a class="reference external" href="http://citeseer.ist.psu.edu/henderson02accurate.html">Accurate Garbage Collection in an Uncooperative Environment</a></p>
</div>
</div>


          </div>
      </div>
      <div class="clearer"></div>
    </div>
    <div class="related">
      <h3>Navigation</h3>
      <ul>
        <li class="right" style="margin-right: 10px">
          <a href="genindex.html" title="General Index"
             >index</a></li>
        <li class="right" >
          <a href="WritingAnLLVMPass.html" title="Writing an LLVM Pass"
             >next</a> |</li>
        <li class="right" >
          <a href="HowToUseInstrMappings.html" title="How To Use Instruction Mappings"
             >previous</a> |</li>
  <li><a href="http://llvm.org/">LLVM Home</a>&nbsp;|&nbsp;</li>
  <li><a href="index.html">Documentation</a>&raquo;</li>
 
      </ul>
    </div>
    <div class="footer">
        &copy; Copyright 2003-2013, LLVM Project.
      Last updated on 2014-03-05.
      Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.1.
    </div>
  </body>
</html>