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
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
|
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>DWARF Extensions For Heterogeneous Debugging — LLVM 13 documentation</title>
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
<link rel="stylesheet" href="_static/llvm-theme.css" type="text/css" />
<script id="documentation_options" data-url_root="./" src="_static/documentation_options.js"></script>
<script src="_static/jquery.js"></script>
<script src="_static/underscore.js"></script>
<script src="_static/doctools.js"></script>
<link rel="index" title="Index" href="genindex.html" />
<link rel="search" title="Search" href="search.html" />
<link rel="next" title="Benchmarking tips" href="Benchmarking.html" />
<link rel="prev" title="AMDGPU Instructions Notation" href="AMDGPUInstructionNotation.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" role="navigation" aria-label="related navigation">
<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="Benchmarking.html" title="Benchmarking tips"
accesskey="N">next</a> |</li>
<li class="right" >
<a href="AMDGPUInstructionNotation.html" title="AMDGPU Instructions Notation"
accesskey="P">previous</a> |</li>
<li><a href="https://llvm.org/">LLVM Home</a> | </li>
<li><a href="index.html">Documentation</a>»</li>
<li class="nav-item nav-item-1"><a href="UserGuides.html" >User Guides</a> »</li>
<li class="nav-item nav-item-2"><a href="AMDGPUUsage.html" accesskey="U">User Guide for AMDGPU Backend</a> »</li>
<li class="nav-item nav-item-this"><a href="">DWARF Extensions For Heterogeneous Debugging</a></li>
</ul>
</div>
<div class="sphinxsidebar" role="navigation" aria-label="main navigation">
<div class="sphinxsidebarwrapper">
<h3>Documentation</h3>
<ul class="want-points">
<li><a href="https://llvm.org/docs/GettingStartedTutorials.html">Getting Started/Tutorials</a></li>
<li><a href="https://llvm.org/docs/UserGuides.html">User Guides</a></li>
<li><a href="https://llvm.org/docs/Reference.html">Reference</a></li>
</ul>
<h3>Getting Involved</h3>
<ul class="want-points">
<li><a href="https://llvm.org/docs/Contributing.html">Contributing to LLVM</a></li>
<li><a href="https://llvm.org/docs/HowToSubmitABug.html">Submitting Bug Reports</a></li>
<li><a href="https://llvm.org/docs/GettingInvolved.html#mailing-lists">Mailing Lists</a></li>
<li><a href="https://llvm.org/docs/GettingInvolved.html#irc">IRC</a></li>
<li><a href="https://llvm.org/docs/GettingInvolved.html#meetups-and-social-events">Meetups and Social Events</a></li>
</ul>
<h3>Additional Links</h3>
<ul class="want-points">
<li><a href="https://llvm.org/docs/FAQ.html">FAQ</a></li>
<li><a href="https://llvm.org/docs/Lexicon.html">Glossary</a></li>
<li><a href="https://llvm.org/pubs">Publications</a></li>
<li><a href="https://github.com/llvm/llvm-project//">Github Repository</a></li>
</ul>
<div role="note" aria-label="source link">
<h3>This Page</h3>
<ul class="this-page-menu">
<li><a href="_sources/AMDGPUDwarfExtensionsForHeterogeneousDebugging.rst.txt"
rel="nofollow">Show Source</a></li>
</ul>
</div>
<div id="searchbox" style="display: none" role="search">
<h3 id="searchlabel">Quick search</h3>
<div class="searchformwrapper">
<form class="search" action="search.html" method="get">
<input type="text" name="q" aria-labelledby="searchlabel" />
<input type="submit" value="Go" />
</form>
</div>
</div>
<script>$('#searchbox').show(0);</script>
</div>
</div>
<div class="document">
<div class="documentwrapper">
<div class="bodywrapper">
<div class="body" role="main">
<div class="section" id="dwarf-extensions-for-heterogeneous-debugging">
<span id="amdgpu-dwarf-extensions-for-heterogeneous-debugging"></span><h1>DWARF Extensions For Heterogeneous Debugging<a class="headerlink" href="#dwarf-extensions-for-heterogeneous-debugging" title="Permalink to this headline">¶</a></h1>
<div class="contents local topic" id="contents">
<ul class="simple">
<li><p><a class="reference internal" href="#introduction" id="id4">Introduction</a></p></li>
<li><p><a class="reference internal" href="#motivation" id="id5">Motivation</a></p></li>
<li><p><a class="reference internal" href="#changes-relative-to-dwarf-version-5" id="id6">Changes Relative to DWARF Version 5</a></p>
<ul>
<li><p><a class="reference internal" href="#general-description" id="id7">General Description</a></p>
<ul>
<li><p><a class="reference internal" href="#attribute-types" id="id8">Attribute Types</a></p></li>
<li><p><a class="reference internal" href="#dwarf-expressions" id="id9">DWARF Expressions</a></p>
<ul>
<li><p><a class="reference internal" href="#dwarf-expression-evaluation-context" id="id10">DWARF Expression Evaluation Context</a></p></li>
<li><p><a class="reference internal" href="#dwarf-expression-value" id="id11">DWARF Expression Value</a></p></li>
<li><p><a class="reference internal" href="#dwarf-location-description" id="id12">DWARF Location Description</a></p></li>
<li><p><a class="reference internal" href="#dwarf-operation-expressions" id="id13">DWARF Operation Expressions</a></p>
<ul>
<li><p><a class="reference internal" href="#stack-operations" id="id14">Stack Operations</a></p></li>
<li><p><a class="reference internal" href="#control-flow-operations" id="id15">Control Flow Operations</a></p></li>
<li><p><a class="reference internal" href="#value-operations" id="id16">Value Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#literal-operations" id="id17">Literal Operations</a></p></li>
<li><p><a class="reference internal" href="#arithmetic-and-logical-operations" id="id18">Arithmetic and Logical Operations</a></p></li>
<li><p><a class="reference internal" href="#type-conversion-operations" id="id19">Type Conversion Operations</a></p></li>
<li><p><a class="reference internal" href="#special-value-operations" id="id20">Special Value Operations</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#location-description-operations" id="id21">Location Description Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#general-location-description-operations" id="id22">General Location Description Operations</a></p></li>
<li><p><a class="reference internal" href="#undefined-location-description-operations" id="id23">Undefined Location Description Operations</a></p></li>
<li><p><a class="reference internal" href="#memory-location-description-operations" id="id24">Memory Location Description Operations</a></p></li>
<li><p><a class="reference internal" href="#register-location-description-operations" id="id25">Register Location Description Operations</a></p></li>
<li><p><a class="reference internal" href="#implicit-location-description-operations" id="id26">Implicit Location Description Operations</a></p></li>
<li><p><a class="reference internal" href="#composite-location-description-operations" id="id27">Composite Location Description Operations</a></p></li>
</ul>
</li>
</ul>
</li>
<li><p><a class="reference internal" href="#dwarf-location-list-expressions" id="id28">DWARF Location List Expressions</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#segmented-addresses" id="id29">Segmented Addresses</a></p></li>
<li><p><a class="reference internal" href="#debugging-information-entry-attributes" id="id30">Debugging Information Entry Attributes</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#program-scope-entities" id="id31">Program Scope Entities</a></p>
<ul>
<li><p><a class="reference internal" href="#unit-entities" id="id32">Unit Entities</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#other-debugger-information" id="id33">Other Debugger Information</a></p>
<ul>
<li><p><a class="reference internal" href="#accelerated-access" id="id34">Accelerated Access</a></p>
<ul>
<li><p><a class="reference internal" href="#lookup-by-name" id="id35">Lookup By Name</a></p>
<ul>
<li><p><a class="reference internal" href="#contents-of-the-name-index" id="id36">Contents of the Name Index</a></p></li>
<li><p><a class="reference internal" href="#data-representation-of-the-name-index" id="id37">Data Representation of the Name Index</a></p>
<ul>
<li><p><a class="reference internal" href="#section-header" id="id38">Section Header</a></p></li>
</ul>
</li>
</ul>
</li>
</ul>
</li>
<li><p><a class="reference internal" href="#line-number-information" id="id39">Line Number Information</a></p>
<ul>
<li><p><a class="reference internal" href="#the-line-number-program-header" id="id40">The Line Number Program Header</a></p>
<ul>
<li><p><a class="reference internal" href="#standard-content-descriptions" id="id41">Standard Content Descriptions</a></p></li>
</ul>
</li>
</ul>
</li>
<li><p><a class="reference internal" href="#call-frame-information" id="id42">Call Frame Information</a></p>
<ul>
<li><p><a class="reference internal" href="#structure-of-call-frame-information" id="id43">Structure of Call Frame Information</a></p></li>
<li><p><a class="reference internal" href="#call-frame-instructions" id="id44">Call Frame Instructions</a></p>
<ul>
<li><p><a class="reference internal" href="#row-creation-instructions" id="id45">Row Creation Instructions</a></p></li>
<li><p><a class="reference internal" href="#cfa-definition-instructions" id="id46">CFA Definition Instructions</a></p></li>
<li><p><a class="reference internal" href="#register-rule-instructions" id="id47">Register Rule Instructions</a></p></li>
<li><p><a class="reference internal" href="#row-state-instructions" id="id48">Row State Instructions</a></p></li>
<li><p><a class="reference internal" href="#padding-instruction" id="id49">Padding Instruction</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#call-frame-instruction-usage" id="id50">Call Frame Instruction Usage</a></p></li>
<li><p><a class="reference internal" href="#call-frame-calling-address" id="id51">Call Frame Calling Address</a></p></li>
</ul>
</li>
</ul>
</li>
<li><p><a class="reference internal" href="#data-representation" id="id52">Data Representation</a></p>
<ul>
<li><p><a class="reference internal" href="#bit-and-64-bit-dwarf-formats" id="id53">32-Bit and 64-Bit DWARF Formats</a></p></li>
<li><p><a class="reference internal" href="#format-of-debugging-information" id="id54">Format of Debugging Information</a></p>
<ul>
<li><p><a class="reference internal" href="#attribute-encodings" id="id55">Attribute Encodings</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#id1" id="id56">DWARF Expressions</a></p>
<ul>
<li><p><a class="reference internal" href="#operation-expressions" id="id57">Operation Expressions</a></p></li>
<li><p><a class="reference internal" href="#location-list-expressions" id="id58">Location List Expressions</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#source-languages" id="id59">Source Languages</a></p></li>
<li><p><a class="reference internal" href="#address-class-and-address-space-encodings" id="id60">Address Class and Address Space Encodings</a></p></li>
<li><p><a class="reference internal" href="#id2" id="id61">Line Number Information</a></p></li>
<li><p><a class="reference internal" href="#id3" id="id62">Call Frame Information</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#attributes-by-tag-value-informative" id="id63">Attributes by Tag Value (Informative)</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#examples" id="id64">Examples</a></p></li>
<li><p><a class="reference internal" href="#references" id="id65">References</a></p></li>
</ul>
</div>
<div class="admonition warning">
<p class="admonition-title">Warning</p>
<p>This document describes <strong>provisional extensions</strong> to DWARF Version 5
[<a class="reference internal" href="#amdgpu-dwarf-dwarf"><span class="std std-ref">DWARF</span></a>] to support heterogeneous debugging. It is
not currently fully implemented and is subject to change.</p>
</div>
<div class="section" id="introduction">
<span id="amdgpu-dwarf-introduction"></span><h2><a class="toc-backref" href="#id4">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
<p>AMD [<a class="reference internal" href="#amdgpu-dwarf-amd"><span class="std std-ref">AMD</span></a>] has been working on supporting heterogeneous
computing through the AMD Radeon Open Compute Platform (ROCm) [<a class="reference internal" href="#amdgpu-dwarf-amd-rocm"><span class="std std-ref">AMD-ROCm</span></a>]. A heterogeneous computing program can be written in a
high level language such as C++ or Fortran with OpenMP pragmas, OpenCL, or HIP
(a portable C++ programming environment for heterogeneous computing [<a class="reference internal" href="#amdgpu-dwarf-hip"><span class="std std-ref">HIP</span></a>]). A heterogeneous compiler and runtime allows a program to
execute on multiple devices within the same native process. Devices could
include CPUs, GPUs, DSPs, FPGAs, or other special purpose accelerators.
Currently HIP programs execute on systems with CPUs and GPUs.</p>
<p>ROCm is fully open sourced and includes contributions to open source projects
such as LLVM for compilation [<a class="reference internal" href="#amdgpu-dwarf-llvm"><span class="std std-ref">LLVM</span></a>] and GDB for
debugging [<a class="reference internal" href="#amdgpu-dwarf-gdb"><span class="std std-ref">GDB</span></a>], as well as collaboration with other
third party projects such as the GCC compiler [<a class="reference internal" href="#amdgpu-dwarf-gcc"><span class="std std-ref">GCC</span></a>]
and the Perforce TotalView HPC debugger [<a class="reference internal" href="#amdgpu-dwarf-perforce-totalview"><span class="std std-ref">Perforce-TotalView</span></a>].</p>
<p>To support debugging heterogeneous programs several features that are not
provided by current DWARF Version 5 [<a class="reference internal" href="#amdgpu-dwarf-dwarf"><span class="std std-ref">DWARF</span></a>] have
been identified. This document contains a collection of extensions to address
providing those features.</p>
<p>The <a class="reference internal" href="#amdgpu-dwarf-motivation"><span class="std std-ref">Motivation</span></a> section describes the issues that are being
addressed for heterogeneous computing. That is followed by the
<a class="reference internal" href="#amdgpu-dwarf-changes-relative-to-dwarf-version-5"><span class="std std-ref">Changes Relative to DWARF Version 5</span></a> section containing the
textual changes for the extensions relative to the DWARF Version 5 standard.
Then there is an <a class="reference internal" href="#amdgpu-dwarf-examples"><span class="std std-ref">Examples</span></a> section that links to the AMD GPU
specific usage of the extensions that includes an example. Finally, there is a
<a class="reference internal" href="#amdgpu-dwarf-references"><span class="std std-ref">References</span></a> section. There are a number of notes included
that raise open questions, or provide alternative approaches considered. The
extensions seek to be general in nature and backwards compatible with DWARF
Version 5. The goal is to be applicable to meeting the needs of any
heterogeneous system and not be vendor or architecture specific.</p>
<p>A fundamental aspect of the extensions is that it allows DWARF expression
location descriptions as stack elements. The extensions are based on DWARF
Version 5 and maintains compatibility with DWARF Version 5. After attempting
several alternatives, the current thinking is that such extensions to DWARF
Version 5 are the simplest and cleanest ways to support debugging optimized GPU
code. It also appears to be generally useful and may be able to address other
reported DWARF issues, as well as being helpful in providing better optimization
support for non-GPU code.</p>
<p>General feedback on these extensions is sought, together with suggestions on how
to clarify, simplify, or organize them. If their is general interest then some
or all of these extensions could be submitted as future DWARF proposals.</p>
<p>We are in the process of modifying LLVM and GDB to support these extensions
which is providing experience and insights. We plan to upstream the changes to
those projects for any final form of the extensions.</p>
<p>The author very much appreciates the input provided so far by many others which
has been incorporated into this current version.</p>
</div>
<div class="section" id="motivation">
<span id="amdgpu-dwarf-motivation"></span><h2><a class="toc-backref" href="#id5">Motivation</a><a class="headerlink" href="#motivation" title="Permalink to this headline">¶</a></h2>
<p>This document presents a set of backwards compatible extensions to DWARF Version
5 [<a class="reference internal" href="#amdgpu-dwarf-dwarf"><span class="std std-ref">DWARF</span></a>] to support heterogeneous debugging.</p>
<p>The remainder of this section provides motivation for each extension in
terms of heterogeneous debugging on commercially available AMD GPU hardware
(AMDGPU). The goal is to add support to the AMD [<a class="reference internal" href="#amdgpu-dwarf-amd"><span class="std std-ref">AMD</span></a>]
open source Radeon Open Compute Platform (ROCm) [<a class="reference internal" href="#amdgpu-dwarf-amd-rocm"><span class="std std-ref">AMD-ROCm</span></a>] which is an implementation of the industry standard
for heterogeneous computing devices defined by the Heterogeneous System
Architecture (HSA) Foundation [<a class="reference internal" href="#amdgpu-dwarf-hsa"><span class="std std-ref">HSA</span></a>]. ROCm includes the
LLVM compiler [<a class="reference internal" href="#amdgpu-dwarf-llvm"><span class="std std-ref">LLVM</span></a>] with upstreamed support for
AMDGPU [<a class="reference internal" href="#amdgpu-dwarf-amdgpu-llvm"><span class="std std-ref">AMDGPU-LLVM</span></a>]. The goal is to also add
the GDB debugger [<a class="reference internal" href="#amdgpu-dwarf-gdb"><span class="std std-ref">GDB</span></a>] with upstreamed support for
AMDGPU [<a class="reference internal" href="#amdgpu-dwarf-amd-rocgdb"><span class="std std-ref">AMD-ROCgdb</span></a>]. In addition, the goal is
to work with third parties to enable support for AMDGPU debugging in the GCC
compiler [<a class="reference internal" href="#amdgpu-dwarf-gcc"><span class="std std-ref">GCC</span></a>] and the Perforce TotalView HPC debugger
[<a class="reference internal" href="#amdgpu-dwarf-perforce-totalview"><span class="std std-ref">Perforce-TotalView</span></a>].</p>
<p>However, the extensions are intended to be vendor and architecture neutral. They
are believed to apply to other heterogeneous hardware devices including GPUs,
DSPs, FPGAs, and other specialized hardware. These collectively include similar
characteristics and requirements as AMDGPU devices. Some of the extension can
also apply to traditional CPU hardware that supports large vector registers.
Compilers can map source languages and extensions that describe large scale
parallel execution onto the lanes of the vector registers. This is common in
programming languages used in ML and HPC. The extensions also include improved
support for optimized code on any architecture. Some of the generalizations may
also benefit other issues that have been raised.</p>
<p>The extensions have evolved through collaboration with many individuals and
active prototyping within the GDB debugger and LLVM compiler. Input has also
been very much appreciated from the developers working on the Perforce TotalView
HPC Debugger and GCC compiler.</p>
<p>The AMDGPU has several features that require additional DWARF functionality in
order to support optimized code.</p>
<p>AMDGPU optimized code may spill vector registers to non-global address space
memory, and this spilling may be done only for lanes that are active on entry
to the subprogram. To support this, a location description that can be created
as a masked select is required. See <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_select_bit_piece</span></code>.</p>
<p>Since the active lane mask may be held in a register, a way to get the value
of a register on entry to a subprogram is required. To support this an
operation that returns the caller value of a register as specified by the Call
Frame Information (CFI) is required. See <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_call_frame_entry_reg</span></code>
and <a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>.</p>
<p>Current DWARF uses an empty expression to indicate an undefined location
description. Since the masked select composite location description operation
takes more than one location description, it is necessary to have an explicit
way to specify an undefined location description. Otherwise it is not possible
to specify that a particular one of the input location descriptions is
undefined. See <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_undefined</span></code>.</p>
<p>CFI describes restoring callee saved registers that are spilled. Currently CFI
only allows a location description that is a register, memory address, or
implicit location description. AMDGPU optimized code may spill scalar
registers into portions of vector registers. This requires extending CFI to
allow any location description. See
<a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>.</p>
<p>The vector registers of the AMDGPU are represented as their full wavefront
size, meaning the wavefront size times the dword size. This reflects the
actual hardware and allows the compiler to generate DWARF for languages that
map a thread to the complete wavefront. It also allows more efficient DWARF to
be generated to describe the CFI as only a single expression is required for
the whole vector register, rather than a separate expression for each lane’s
dword of the vector register. It also allows the compiler to produce DWARF
that indexes the vector register if it spills scalar registers into portions
of a vector register.</p>
<p>Since DWARF stack value entries have a base type and AMDGPU registers are a
vector of dwords, the ability to specify that a base type is a vector is
required. See <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_vector_size</span></code>.</p>
<p>If the source language is mapped onto the AMDGPU wavefronts in a SIMT manner,
then the variable DWARF location expressions must compute the location for a
single lane of the wavefront. Therefore, a DWARF operation is required to denote
the current lane, much like <code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> denotes the current
object. The <code class="docutils literal notranslate"><span class="pre">DW_OP_*piece</span></code> operations only allow literal indices. Therefore, a
way to use a computed offset of an arbitrary location description (such as a
vector register) is required. See <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_push_lane</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset_uconst</span></code>, and
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_bit_offset</span></code>.</p>
<p>If the source language is mapped onto the AMDGPU wavefronts in a SIMT manner
the compiler can use the AMDGPU execution mask register to control which lanes
are active. To describe the conceptual location of non-active lanes a DWARF
expression is needed that can compute a per lane PC. For efficiency, this is
done for the wavefront as a whole. This expression benefits by having a masked
select composite location description operation. This requires an attribute
for source location of each lane. The AMDGPU may update the execution mask for
whole wavefront operations and so needs an attribute that computes the current
active lane mask. See <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_select_bit_piece</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_extend</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code>, and <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_active_lane</span></code>.</p>
<p>AMDGPU needs to be able to describe addresses that are in different kinds of
memory. Optimized code may need to describe a variable that resides in pieces
that are in different kinds of storage which may include parts of registers,
memory that is in a mixture of memory kinds, implicit values, or be undefined.
DWARF has the concept of segment addresses. However, the segment cannot be
specified within a DWARF expression, which is only able to specify the offset
portion of a segment address. The segment index is only provided by the entity
that specifies the DWARF expression. Therefore, the segment index is a
property that can only be put on complete objects, such as a variable. That
makes it only suitable for describing an entity (such as variable or
subprogram code) that is in a single kind of memory. Therefore, AMDGPU uses
the DWARF concept of address spaces. For example, a variable may be allocated
in a register that is partially spilled to the call stack which is in the
private address space, and partially spilled to the local address space.</p>
<p>DWARF uses the concept of an address in many expression operations but does not
define how it relates to address spaces. For example,
<code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> pushes the address of an object. Other contexts
implicitly push an address on the stack before evaluating an expression. For
example, the <code class="docutils literal notranslate"><span class="pre">DW_AT_use_location</span></code> attribute of the
<code class="docutils literal notranslate"><span class="pre">DW_TAG_ptr_to_member_type</span></code>. The expression that uses the address needs to
do so in a general way and not need to be dependent on the address space of
the address. For example, a pointer to member value may want to be applied to
an object that may reside in any address space.</p>
<p>The number of registers and the cost of memory operations is much higher for
AMDGPU than a typical CPU. The compiler attempts to optimize whole variables
and arrays into registers. Currently DWARF only allows
<code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> and related operations to work with a global
memory location. To support AMDGPU optimized code it is required to generalize
DWARF to allow any location description to be used. This allows registers, or
composite location descriptions that may be a mixture of memory, registers, or
even implicit values.</p>
<p>DWARF Version 5 does not allow location descriptions to be entries on the
DWARF stack. They can only be the final result of the evaluation of a DWARF
expression. However, by allowing a location description to be a first-class
entry on the DWARF stack it becomes possible to compose expressions containing
both values and location descriptions naturally. It allows objects to be
located in any kind of memory address space, in registers, be implicit values,
be undefined, or a composite of any of these. By extending DWARF carefully,
all existing DWARF expressions can retain their current semantic meaning.
DWARF has implicit conversions that convert from a value that represents an
address in the default address space to a memory location description. This
can be extended to allow a default address space memory location description
to be implicitly converted back to its address value. This allows all DWARF
Version 5 expressions to retain their same meaning, while adding the ability
to explicitly create memory location descriptions in non-default address
spaces and generalizing the power of composite location descriptions to any
kind of location description. See <a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>.</p>
<p>To allow composition of composite location descriptions, an explicit operation
that indicates the end of the definition of a composite location description
is required. This can be implied if the end of a DWARF expression is reached,
allowing current DWARF expressions to remain legal. See
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_piece_end</span></code>.</p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_OP_plus</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_OP_minus</span></code> can be defined to operate on a memory
location description in the default target architecture specific address space
and a generic type value to produce an updated memory location description. This
allows them to continue to be used to offset an address. To generalize
offsetting to any location description, including location descriptions that
describe when bytes are in registers, are implicit, or a composite of these, the
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset_uconst</span></code>, and
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_bit_offset</span></code> offset operations are added. Unlike <code class="docutils literal notranslate"><span class="pre">DW_OP_plus</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_OP_plus_uconst</span></code>, and <code class="docutils literal notranslate"><span class="pre">DW_OP_minus</span></code> arithmetic operations, these do not
define that integer overflow causes wrap-around. The offset operations can
operate on location storage of any size. For example, implicit location storage
could be any number of bits in size. It is simpler to define offsets that exceed
the size of the location storage as being an evaluation error, than having to
force an implementation to support potentially infinite precision offsets to
allow it to correctly track a series of positive and negative offsets that may
transiently overflow or underflow, but end up in range. This is simple for the
arithmetic operations as they are defined in terms of two’s compliment
arithmetic on a base type of a fixed size.</p>
<p>Having the offset operations allows <code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> to push a
location description that may be in a register, or be an implicit value, and the
DWARF expression of <code class="docutils literal notranslate"><span class="pre">DW_TAG_ptr_to_member_type</span></code> can contain them to offset
within it. <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_bit_offset</span></code> generalizes DWARF to work with bit fields
which is not possible in DWARF Version 5.</p>
<p>The DWARF <code class="docutils literal notranslate"><span class="pre">DW_OP_xderef*</span></code> operations allow a value to be converted into an
address of a specified address space which is then read. But it provides no
way to create a memory location description for an address in the non-default
address space. For example, AMDGPU variables can be allocated in the local
address space at a fixed address. It is required to have an operation to
create an address in a specific address space that can be used to define the
location description of the variable. Defining this operation to produce a
location description allows the size of addresses in an address space to be
larger than the generic type. See <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code>.</p>
<p>If the <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code> operation had to produce a value
that can be implicitly converted to a memory location description, then it
would be limited to the size of the generic type which matches the size of the
default address space. Its value would be undefined and likely not match any
value in the actual program. By making the result a location description, it
allows a consumer great freedom in how it implements it. The implicit
conversion back to a value can be limited only to the default address space to
maintain compatibility with DWARF Version 5. For other address spaces the
producer can use the new operations that explicitly specify the address space.</p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_breg*</span></code> treats the register as containing an address in the default
address space. It is required to be able to specify the address space of the
register value. See <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_bregx</span></code>.</p>
<p>Similarly, <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> treats its implicit pointer value as
being in the default address space. It is required to be able to specify the
address space of the pointer value. See
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code>.</p>
<p>Almost all uses of addresses in DWARF are limited to defining location
descriptions, or to be dereferenced to read memory. The exception is
<code class="docutils literal notranslate"><span class="pre">DW_CFA_val_offset</span></code> which uses the address to set the value of a register.
By defining the CFA DWARF expression as being a memory location description,
it can maintain what address space it is, and that can be used to convert the
offset address back to an address in that address space. See
<a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>.</p>
<p>This approach allows all existing DWARF to have the identical semantics. It
allows the compiler to explicitly specify the address space it is using. For
example, a compiler could choose to access private memory in a swizzled manner
when mapping a source language to a wavefront in a SIMT manner, or to access
it in an unswizzled manner if mapping the same language with the wavefront
being the thread. It also allows the compiler to mix the address space it uses
to access private memory. For example, for SIMT it can still spill entire
vector registers in an unswizzled manner, while using a swizzled private
memory for SIMT variable access. This approach allows memory location
descriptions for different address spaces to be combined using the regular
<code class="docutils literal notranslate"><span class="pre">DW_OP_*piece</span></code> operations.</p>
<p>Location descriptions are an abstraction of storage, they give freedom to the
consumer on how to implement them. They allow the address space to encode lane
information so they can be used to read memory with only the memory
description and no extra arguments. The same set of operations can operate on
locations independent of their kind of storage. The <code class="docutils literal notranslate"><span class="pre">DW_OP_deref*</span></code> therefore
can be used on any storage kind. <code class="docutils literal notranslate"><span class="pre">DW_OP_xderef*</span></code> is unnecessary, except to
become a more compact way to convert a non-default address space address
followed by dereferencing it.</p>
<p>In DWARF Version 5 a location description is defined as a single location
description or a location list. A location list is defined as either
effectively an undefined location description or as one or more single
location descriptions to describe an object with multiple places. The
<code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> operations can put a
location description on the stack. Furthermore, debugger information entry
attributes such as <code class="docutils literal notranslate"><span class="pre">DW_AT_data_member_location</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_AT_use_location</span></code>, and
<code class="docutils literal notranslate"><span class="pre">DW_AT_vtable_elem_location</span></code> are defined as pushing a location description
on the expression stack before evaluating the expression. However, DWARF
Version 5 only allows the stack to contain values and so only a single memory
address can be on the stack which makes these incapable of handling location
descriptions with multiple places, or places other than memory. Since these
extensions allow the stack to contain location descriptions, the operations are
generalized to support location descriptions that can have multiple places.
This is backwards compatible with DWARF Version 5 and allows objects with
multiple places to be supported. For example, the expression that describes
how to access the field of an object can be evaluated with a location
description that has multiple places and will result in a location description
with multiple places as expected. With this change, the separate DWARF Version
5 sections that described DWARF expressions and location lists have been
unified into a single section that describes DWARF expressions in general.
This unification seems to be a natural consequence and a necessity of allowing
location descriptions to be part of the evaluation stack.</p>
<p>For those familiar with the definition of location descriptions in DWARF Version
5, the definitions in these extensions are presented differently, but does
in fact define the same concept with the same fundamental semantics. However,
it does so in a way that allows the concept to extend to support address
spaces, bit addressing, the ability for composite location descriptions to be
composed of any kind of location description, and the ability to support
objects located at multiple places. Collectively these changes expand the set
of processors that can be supported and improves support for optimized code.</p>
<p>Several approaches were considered, and the one presented appears to be the
cleanest and offers the greatest improvement of DWARF’s ability to support
optimized code. Examining the GDB debugger and LLVM compiler, it appears only
to require modest changes as they both already have to support general use of
location descriptions. It is anticipated that will also be the case for other
debuggers and compilers.</p>
<p>As an experiment, GDB was modified to evaluate DWARF Version 5 expressions
with location descriptions as stack entries and implicit conversions. All GDB
tests have passed, except one that turned out to be an invalid test by DWARF
Version 5 rules. The code in GDB actually became simpler as all evaluation was
on the stack and there was no longer a need to maintain a separate structure
for the location description result. This gives confidence of the backwards
compatibility.</p>
<p>Since the AMDGPU supports languages such as OpenCL [<a class="reference internal" href="#amdgpu-dwarf-opencl"><span class="std std-ref">OpenCL</span></a>], there is a need to define source language address
classes so they can be used in a consistent way by consumers. It would also be
desirable to add support for using them in defining language types rather than
the current target architecture specific address spaces. See
<a class="reference internal" href="#amdgpu-dwarf-segment-addresses"><span class="std std-ref">Segmented Addresses</span></a>.</p>
<p>A <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_augmentation</span></code> attribute is added to a compilation unit
debugger information entry to indicate that there is additional target
architecture specific information in the debugging information entries of that
compilation unit. This allows a consumer to know what extensions are present
in the debugger information entries as is possible with the augmentation
string of other sections. The format that should be used for the augmentation
string in the lookup by name table and CFI Common Information Entry is also
recommended to allow a consumer to parse the string when it contains
information from multiple vendors.</p>
<p>The AMDGPU supports programming languages that include online compilation
where the source text may be created at runtime. Therefore, a way to embed the
source text in the debug information is required. For example, the OpenCL
language runtime supports online compilation. See
<a class="reference internal" href="#amdgpu-dwarf-line-number-information"><span class="std std-ref">Line Number Information</span></a>.</p>
<p>Support to allow MD5 checksums to be optionally present in the line table is
added. This allows linking together compilation units where some have MD5
checksums and some do not. In DWARF Version 5 the file timestamp and file size
can be optional, but if the MD5 checksum is present it must be valid for all
files. See <a class="reference internal" href="#amdgpu-dwarf-line-number-information"><span class="std std-ref">Line Number Information</span></a>.</p>
<p>Support is added for the HIP programming language [<a class="reference internal" href="#amdgpu-dwarf-hip"><span class="std std-ref">HIP</span></a>] which is supported by the AMDGPU. See
<a class="reference internal" href="#amdgpu-dwarf-language-names"><span class="std std-ref">Unit Entities</span></a>.</p>
<p>The following sections provide the definitions for the additional operations,
as well as clarifying how existing expression operations, CFI operations, and
attributes behave with respect to generalized location descriptions that
support address spaces and location descriptions that support multiple places.
It has been defined such that it is backwards compatible with DWARF Version 5.
The definitions are intended to fully define well-formed DWARF in a consistent
style based on the DWARF Version 5 specification. Non-normative text is shown
in <em>italics</em>.</p>
<p>The names for the new operations, attributes, and constants include “<code class="docutils literal notranslate"><span class="pre">LLVM</span></code>” and are encoded with vendor specific codes so these extensions can
be implemented as an LLVM vendor extension to DWARF Version 5. If accepted these
names would not include the “<code class="docutils literal notranslate"><span class="pre">LLVM</span></code>” and would not use encodings in the
vendor range.</p>
<p>The extensions are described in
<a class="reference internal" href="#amdgpu-dwarf-changes-relative-to-dwarf-version-5"><span class="std std-ref">Changes Relative to DWARF Version 5</span></a> and are
organized to follow the section ordering of DWARF Version 5. It includes notes
to indicate the corresponding DWARF Version 5 sections to which they pertain.
Other notes describe additional changes that may be worth considering, and to
raise questions.</p>
</div>
<div class="section" id="changes-relative-to-dwarf-version-5">
<span id="amdgpu-dwarf-changes-relative-to-dwarf-version-5"></span><h2><a class="toc-backref" href="#id6">Changes Relative to DWARF Version 5</a><a class="headerlink" href="#changes-relative-to-dwarf-version-5" title="Permalink to this headline">¶</a></h2>
<div class="section" id="general-description">
<h3><a class="toc-backref" href="#id7">General Description</a><a class="headerlink" href="#general-description" title="Permalink to this headline">¶</a></h3>
<div class="section" id="attribute-types">
<h4><a class="toc-backref" href="#id8">Attribute Types</a><a class="headerlink" href="#attribute-types" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 2.2 and Table 2.2.</p>
</div>
<p>The following table provides the additional attributes. See
<a class="reference internal" href="#amdgpu-dwarf-debugging-information-entry-attributes"><span class="std std-ref">Debugging Information Entry Attributes</span></a>.</p>
<table class="docutils align-default" id="amdgpu-dwarf-attribute-names-table">
<caption><span class="caption-text">Attribute names</span><a class="headerlink" href="#amdgpu-dwarf-attribute-names-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 43%" />
<col style="width: 57%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Attribute</p></th>
<th class="head"><p>Usage</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_active_lane</span></code></p></td>
<td><p>SIMD or SIMT active lanes</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_augmentation</span></code></p></td>
<td><p>Compilation unit augmentation string</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code></p></td>
<td><p>SIMD or SIMT lane program location</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lanes</span></code></p></td>
<td><p>SIMD or SIMT thread lane count</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_vector_size</span></code></p></td>
<td><p>Base type vector size</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="dwarf-expressions">
<span id="amdgpu-dwarf-expressions"></span><h4><a class="toc-backref" href="#id9">DWARF Expressions</a><a class="headerlink" href="#dwarf-expressions" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This section, and its nested sections, replaces DWARF Version 5 section 2.5
and section 2.6. The new DWARF expression operation extensions are defined as
well as clarifying the extensions to already existing DWARF Version 5
operations. It is based on the text of the existing DWARF Version 5 standard.</p>
</div>
<p>DWARF expressions describe how to compute a value or specify a location.</p>
<p><em>The evaluation of a DWARF expression can provide the location of an object, the
value of an array bound, the length of a dynamic string, the desired value
itself, and so on.</em></p>
<p>If the evaluation of a DWARF expression does not encounter an error, then it can
either result in a value (see <a class="reference internal" href="#amdgpu-dwarf-expression-value"><span class="std std-ref">DWARF Expression Value</span></a>) or a
location description (see <a class="reference internal" href="#amdgpu-dwarf-location-description"><span class="std std-ref">DWARF Location Description</span></a>). When a
DWARF expression is evaluated, it may be specified whether a value or location
description is required as the result kind.</p>
<p>If a result kind is specified, and the result of the evaluation does not match
the specified result kind, then the implicit conversions described in
<a class="reference internal" href="#amdgpu-dwarf-memory-location-description-operations"><span class="std std-ref">Memory Location Description Operations</span></a> are performed if
valid. Otherwise, the DWARF expression is ill-formed.</p>
<p>If the evaluation of a DWARF expression encounters an evaluation error, then the
result is an evaluation error.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Decided to define the concept of an evaluation error. An alternative is to
introduce an undefined value base type in a similar way to location
descriptions having an undefined location description. Then operations that
encounter an evaluation error can return the undefined location description or
value with an undefined base type.</p>
<p>All operations that act on values would return an undefined entity if given an
undefined value. The expression would then always evaluate to completion, and
can be tested to determine if it is an undefined entity.</p>
<p>However, this would add considerable additional complexity and does not match
that GDB throws an exception when these evaluation errors occur.</p>
</div>
<p>If a DWARF expression is ill-formed, then the result is undefined.</p>
<p>The following sections detail the rules for when a DWARF expression is
ill-formed or results in an evaluation error.</p>
<p>A DWARF expression can either be encoded as an operation expression (see
<a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>), or as a location list expression
(see <a class="reference internal" href="#amdgpu-dwarf-location-list-expressions"><span class="std std-ref">DWARF Location List Expressions</span></a>).</p>
<div class="section" id="dwarf-expression-evaluation-context">
<span id="amdgpu-dwarf-expression-evaluation-context"></span><h5><a class="toc-backref" href="#id10">DWARF Expression Evaluation Context</a><a class="headerlink" href="#dwarf-expression-evaluation-context" title="Permalink to this headline">¶</a></h5>
<p>A DWARF expression is evaluated in a context that can include a number of
context elements. If multiple context elements are specified then they must be
self consistent or the result of the evaluation is undefined. The context
elements that can be specified are:</p>
<p><em>A current result kind</em></p>
<blockquote>
<div><p>The kind of result required by the DWARF expression evaluation. If specified
it can be a location description or a value.</p>
</div></blockquote>
<p><em>A current thread</em></p>
<blockquote>
<div><p>The target architecture thread identifier of the source program thread of
execution for which a user presented expression is currently being evaluated.</p>
<p>It is required for operations that are related to target architecture threads.</p>
<p><em>For example, the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_form_tls_address</span></code> <em>operation and</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code> <em>operation when given an address space that
is thread specific.</em></p>
</div></blockquote>
<p><em>A current lane</em></p>
<blockquote>
<div><p>The target architecture lane identifier of the source program thread of
execution for which a user presented expression is currently being evaluated.
This applies to languages that are implemented using a SIMD or SIMT execution
model.</p>
<p>It is required for operations that are related to target architecture lanes.</p>
<p><em>For example, the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_push_lane</span></code> <em>operation and</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code> <em>operation when given an address space that
is lane specific.</em></p>
<p>If specified, it must be consistent with any specified current thread and
current target architecture. It is consistent with a thread if it identifies a
lane of the thread. It is consistent with a target architecture if it is a
valid lane identifier of the target architecture. Otherwise the result is
undefined.</p>
</div></blockquote>
<p><em>A current call frame</em></p>
<blockquote>
<div><p>The target architecture call frame identifier. It identifies a call frame that
corresponds to an active invocation of a subprogram in the current thread. It
is identified by its address on the call stack. The address is referred to as
the Canonical Frame Address (CFA). The call frame information is used to
determine the CFA for the call frames of the current thread’s call stack (see
<a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>).</p>
<p>It is required for operations that specify target architecture registers to
support virtual unwinding of the call stack.</p>
<p><em>For example, the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_*reg*</span></code> <em>operations.</em></p>
<p>If specified, it must be an active call frame in the current thread. If the
current lane is specified, then that lane must have been active on entry to
the call frame (see the <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code> attribute). Otherwise the
result is undefined.</p>
<p>If it is the currently executing call frame, then it is termed the top call
frame.</p>
</div></blockquote>
<p><em>A current program location</em></p>
<blockquote>
<div><p>The target architecture program location corresponding to the current call
frame of the current thread.</p>
<p>The program location of the top call frame is the target architecture program
counter for the current thread. The call frame information is used to obtain
the value of the return address register to determine the program location of
the other call frames (see <a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>).</p>
<p>It is required for the evaluation of location list expressions to select
amongst multiple program location ranges. It is required for operations that
specify target architecture registers to support virtual unwinding of the call
stack (see <a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>).</p>
<p>If specified:</p>
<ul class="simple">
<li><p>If the current lane is not specified:</p>
<ul>
<li><p>If the current call frame is the top call frame, it must be the current
target architecture program location.</p></li>
<li><p>If the current call frame F is not the top call frame, it must be the
program location associated with the call site in the current caller frame
F that invoked the callee frame.</p></li>
</ul>
</li>
<li><p>If the current lane is specified and the architecture program location LPC
computed by the <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code> attribute for the current lane is not
the undefined location description (indicating the lane was not active on
entry to the call frame), it must be LPC.</p></li>
<li><p>Otherwise the result is undefined.</p></li>
</ul>
</div></blockquote>
<p><em>A current compilation unit</em></p>
<blockquote>
<div><p>The compilation unit debug information entry that contains the DWARF expression
being evaluated.</p>
<p>It is required for operations that reference debug information associated with
the same compilation unit, including indicating if such references use the
32-bit or 64-bit DWARF format. It can also provide the default address space
address size if no current target architecture is specified.</p>
<p><em>For example, the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_constx</span></code> <em>and</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_addrx</span></code> <em>operations.</em></p>
<p><em>Note that this compilation unit may not be the same as the compilation unit
determined from the loaded code object corresponding to the current program
location. For example, the evaluation of the expression E associated with a
``DW_AT_location`` attribute of the debug information entry operand of the
``DW_OP_call*`` operations is evaluated with the compilation unit that
contains E and not the one that contains the ``DW_OP_call*`` operation
expression.</em></p>
</div></blockquote>
<p><em>A current target architecture</em></p>
<blockquote>
<div><p>The target architecture.</p>
<p>It is required for operations that specify target architecture specific
entities.</p>
<p><em>For example, target architecture specific entities include DWARF register
identifiers, DWARF lane identifiers, DWARF address space identifiers, the
default address space, and the address space address sizes.</em></p>
<p>If specified:</p>
<ul class="simple">
<li><p>If the current thread is specified, then the current target architecture
must be the same as the target architecture of the current thread.</p></li>
<li><p>If the current compilation unit is specified, then the current target
architecture default address space address size must be the same as he
<code class="docutils literal notranslate"><span class="pre">address_size</span></code> field in the header of the current compilation unit and any
associated entry in the <code class="docutils literal notranslate"><span class="pre">.debug_aranges</span></code> section.</p></li>
<li><p>If the current program location is specified, then the current target
architecture must be the same as the target architecture of any line number
information entry (see <a class="reference internal" href="#amdgpu-dwarf-line-number-information"><span class="std std-ref">Line Number Information</span></a>)
corresponding to the current program location.</p></li>
<li><p>If the current program location is specified, then the current target
architecture default address space address size must be the same as he
<code class="docutils literal notranslate"><span class="pre">address_size</span></code> field in the header of any entry corresponding to the
current program location in the <code class="docutils literal notranslate"><span class="pre">.debug_addr</span></code>, <code class="docutils literal notranslate"><span class="pre">.debug_line</span></code>,
<code class="docutils literal notranslate"><span class="pre">.debug_rnglists</span></code>, <code class="docutils literal notranslate"><span class="pre">.debug_rnglists.dwo</span></code>, <code class="docutils literal notranslate"><span class="pre">.debug_loclists</span></code>, and
<code class="docutils literal notranslate"><span class="pre">.debug_loclists.dwo</span></code> sections.</p></li>
<li><p>Otherwise the result is undefined.</p></li>
</ul>
</div></blockquote>
<p><em>A current object</em></p>
<blockquote>
<div><p>The location description of a program object.</p>
<p>It is required for the <code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> operation.</p>
<p><em>For example, the</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_data_location</span></code> <em>attribute on type debug
information entries specifies the the program object corresponding to a
runtime descriptor as the current object when it evaluates its associated
expression.</em></p>
<p>The result is undefined if the location descriptor is invalid (see
<a class="reference internal" href="#amdgpu-dwarf-location-description"><span class="std std-ref">DWARF Location Description</span></a>).</p>
</div></blockquote>
<p><em>An initial stack</em></p>
<blockquote>
<div><p>This is a list of values or location descriptions that will be pushed on the
operation expression evaluation stack in the order provided before evaluation
of an operation expression starts.</p>
<p>Some debugger information entries have attributes that evaluate their DWARF
expression value with initial stack entries. In all other cases the initial
stack is empty.</p>
<p>The result is undefined if any location descriptors are invalid (see
<a class="reference internal" href="#amdgpu-dwarf-location-description"><span class="std std-ref">DWARF Location Description</span></a>).</p>
</div></blockquote>
<p>If the evaluation requires a context element that is not specified, then the
result of the evaluation is an error.</p>
<p><em>A DWARF expression for the location description may be able to be evaluated
without a thread, lane, call frame, program location, or architecture context.
For example, the location of a global variable may be able to be evaluated
without such context. If the expression evaluates with an error then it may
indicate the variable has been optimized and so requires more context.</em></p>
<p><em>The DWARF expression for call frame information (see
:ref:`amdgpu-dwarf-call-frame-information`) operations are restricted to those
that do not require the compilation unit context to be specified.</em></p>
<p>The DWARF is ill-formed if all the <code class="docutils literal notranslate"><span class="pre">address_size</span></code> fields in the headers of all
the entries in the <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code>, <code class="docutils literal notranslate"><span class="pre">.debug_addr</span></code>, <code class="docutils literal notranslate"><span class="pre">.debug_line</span></code>,
<code class="docutils literal notranslate"><span class="pre">.debug_rnglists</span></code>, <code class="docutils literal notranslate"><span class="pre">.debug_rnglists.dwo</span></code>, <code class="docutils literal notranslate"><span class="pre">.debug_loclists</span></code>, and
<code class="docutils literal notranslate"><span class="pre">.debug_loclists.dwo</span></code> sections corresponding to any given program location do
not match.</p>
</div>
<div class="section" id="dwarf-expression-value">
<span id="amdgpu-dwarf-expression-value"></span><h5><a class="toc-backref" href="#id11">DWARF Expression Value</a><a class="headerlink" href="#dwarf-expression-value" title="Permalink to this headline">¶</a></h5>
<p>A value has a type and a literal value. It can represent a literal value of any
supported base type of the target architecture. The base type specifies the
size, encoding, and endianity of the literal value.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>It may be desirable to add an implicit pointer base type encoding. It would be
used for the type of the value that is produced when the <code class="docutils literal notranslate"><span class="pre">DW_OP_deref*</span></code>
operation retrieves the full contents of an implicit pointer location storage
created by the <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> or
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code> operations. The literal value would
record the debugging information entry and byte displacement specified by the
associated <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> or
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code> operations.</p>
</div>
<p>There is a distinguished base type termed the generic type, which is an integral
type that has the size of an address in the target architecture default address
space, a target architecture defined endianity, and unspecified signedness.</p>
<p><em>The generic type is the same as the unspecified type used for stack operations
defined in DWARF Version 4 and before.</em></p>
<p>An integral type is a base type that has an encoding of <code class="docutils literal notranslate"><span class="pre">DW_ATE_signed</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_ATE_signed_char</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_ATE_unsigned</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_ATE_unsigned_char</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_ATE_boolean</span></code>, or any target architecture defined integral encoding in the
inclusive range <code class="docutils literal notranslate"><span class="pre">DW_ATE_lo_user</span></code> to <code class="docutils literal notranslate"><span class="pre">DW_ATE_hi_user</span></code>.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>It is unclear if <code class="docutils literal notranslate"><span class="pre">DW_ATE_address</span></code> is an integral type. GDB does not seem to
consider it as integral.</p>
</div>
</div>
<div class="section" id="dwarf-location-description">
<span id="amdgpu-dwarf-location-description"></span><h5><a class="toc-backref" href="#id12">DWARF Location Description</a><a class="headerlink" href="#dwarf-location-description" title="Permalink to this headline">¶</a></h5>
<p><em>Debugging information must provide consumers a way to find the location of
program variables, determine the bounds of dynamic arrays and strings, and
possibly to find the base address of a subprogram’s call frame or the return
address of a subprogram. Furthermore, to meet the needs of recent computer
architectures and optimization techniques, debugging information must be able to
describe the location of an object whose location changes over the object’s
lifetime, and may reside at multiple locations simultaneously during parts of an
object’s lifetime.</em></p>
<p>Information about the location of program objects is provided by location
descriptions.</p>
<p>Location descriptions can consist of one or more single location descriptions.</p>
<p>A single location description specifies the location storage that holds a
program object and a position within the location storage where the program
object starts. The position within the location storage is expressed as a bit
offset relative to the start of the location storage.</p>
<p>A location storage is a linear stream of bits that can hold values. Each
location storage has a size in bits and can be accessed using a zero-based bit
offset. The ordering of bits within a location storage uses the bit numbering
and direction conventions that are appropriate to the current language on the
target architecture.</p>
<p>There are five kinds of location storage:</p>
<dl class="simple">
<dt><em>memory location storage</em></dt><dd><p>Corresponds to the target architecture memory address spaces.</p>
</dd>
<dt><em>register location storage</em></dt><dd><p>Corresponds to the target architecture registers.</p>
</dd>
<dt><em>implicit location storage</em></dt><dd><p>Corresponds to fixed values that can only be read.</p>
</dd>
<dt><em>undefined location storage</em></dt><dd><p>Indicates no value is available and therefore cannot be read or written.</p>
</dd>
<dt><em>composite location storage</em></dt><dd><p>Allows a mixture of these where some bits come from one location storage and
some from another location storage, or from disjoint parts of the same
location storage.</p>
</dd>
</dl>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>It may be better to add an implicit pointer location storage kind used by the
<code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code>
operations. It would specify the debugger information entry and byte offset
provided by the operations.</p>
</div>
<p><em>Location descriptions are a language independent representation of addressing
rules. They are created using DWARF operation expressions of arbitrary
complexity. They can be the result of evaluating a debugger information entry
attribute that specifies an operation expression. In this usage they can
describe the location of an object as long as its lifetime is either static or
the same as the lexical block (see DWARF Version 5 section 3.5) that owns it,
and it does not move during its lifetime. They can be the result of evaluating a
debugger information entry attribute that specifies a location list expression.
In this usage they can describe the location of an object that has a limited
lifetime, changes its location during its lifetime, or has multiple locations
over part or all of its lifetime.</em></p>
<p>If a location description has more than one single location description, the
DWARF expression is ill-formed if the object value held in each single location
description’s position within the associated location storage is not the same
value, except for the parts of the value that are uninitialized.</p>
<p><em>A location description that has more than one single location description can
only be created by a location list expression that has overlapping program
location ranges, or certain expression operations that act on a location
description that has more than one single location description. There are no
operation expression operations that can directly create a location description
with more than one single location description.</em></p>
<p><em>A location description with more than one single location description can be
used to describe objects that reside in more than one piece of storage at the
same time. An object may have more than one location as a result of
optimization. For example, a value that is only read may be promoted from memory
to a register for some region of code, but later code may revert to reading the
value from memory as the register may be used for other purposes. For the code
region where the value is in a register, any change to the object value must be
made in both the register and the memory so both regions of code will read the
updated value.</em></p>
<p><em>A consumer of a location description with more than one single location
description can read the object’s value from any of the single location
descriptions (since they all refer to location storage that has the same value),
but must write any changed value to all the single location descriptions.</em></p>
<p>The evaluation of an expression may require context elements to create a
location description. If such a location description is accessed, the storage it
denotes is that associated with the context element values specified when the
location description was created, which may differ from the context at the time
it is accessed.</p>
<p><em>For example, creating a register location description requires the thread
context: the location storage is for the specified register of that thread.
Creating a memory location description for an address space may required a
thread and a lane context: the location storage is the memory associated with
that thread and lane.</em></p>
<p>If any of the context elements required to create a location description change,
the location description becomes invalid and accessing it is undefined.</p>
<p><em>Examples of context that can invalidate a location description are:</em></p>
<ul class="simple">
<li><p><em>The thread context is required and execution causes the thread to terminate.</em></p></li>
<li><p><em>The call frame context is required and further execution causes the call
frame to return to the calling frame.</em></p></li>
<li><p><em>The program location is required and further execution of the thread occurs.
That could change the location list entry or call frame information entry that
applies.</em></p></li>
<li><p><em>An operation uses call frame information:</em></p>
<ul>
<li><p><em>Any of the frames used in the virtual call frame unwinding return.</em></p></li>
<li><p><em>The top call frame is used, the program location is used to select the call
frame information entry, and further execution of the thread occurs.</em></p></li>
</ul>
</li>
</ul>
<p><em>A DWARF expression can be used to compute a location description for an object.
A subsequent DWARF expression evaluation can be given the object location
description as the object context or initial stack context to compute a
component of the object. The final result is undefined if the object location
description becomes invalid between the two expression evaluations.</em></p>
<p>A change of a thread’s program location may not make a location description
invalid, yet may still render it as no longer meaningful. Accessing such a
location description, or using it as the object context or initial stack context
of an expression evaluation, may produce an undefined result.</p>
<p><em>For example, a location description may specify a register that no longer holds
the intended program object after a program location change. One way to avoid
such problems is to recompute location descriptions associated with threads when
their program locations change.</em></p>
</div>
<div class="section" id="dwarf-operation-expressions">
<span id="amdgpu-dwarf-operation-expressions"></span><h5><a class="toc-backref" href="#id13">DWARF Operation Expressions</a><a class="headerlink" href="#dwarf-operation-expressions" title="Permalink to this headline">¶</a></h5>
<p>An operation expression is comprised of a stream of operations, each consisting
of an opcode followed by zero or more operands. The number of operands is
implied by the opcode.</p>
<p>Operations represent a postfix operation on a simple stack machine. Each stack
entry can hold either a value or a location description. Operations can act on
entries on the stack, including adding entries and removing entries. If the kind
of a stack entry does not match the kind required by the operation and is not
implicitly convertible to the required kind (see
<a class="reference internal" href="#amdgpu-dwarf-memory-location-description-operations"><span class="std std-ref">Memory Location Description Operations</span></a>), then the DWARF
operation expression is ill-formed.</p>
<p>Evaluation of an operation expression starts with an empty stack on which the
entries from the initial stack provided by the context are pushed in the order
provided. Then the operations are evaluated, starting with the first operation
of the stream. Evaluation continues until either an operation has an evaluation
error, or until one past the last operation of the stream is reached.</p>
<p>The result of the evaluation is:</p>
<ul>
<li><p>If an operation has an evaluation error, or an operation evaluates an
expression that has an evaluation error, then the result is an evaluation
error.</p></li>
<li><p>If the current result kind specifies a location description, then:</p>
<ul>
<li><p>If the stack is empty, the result is a location description with one
undefined location description.</p>
<p><em>This rule is for backwards compatibility with DWARF Version 5 which has no
explicit operation to create an undefined location description, and uses an
empty operation expression for this purpose.</em></p>
</li>
<li><p>If the top stack entry is a location description, or can be converted
to one (see <a class="reference internal" href="#amdgpu-dwarf-memory-location-description-operations"><span class="std std-ref">Memory Location Description Operations</span></a>),
then the result is that, possibly converted, location description. Any other
entries on the stack are discarded.</p></li>
<li><p>Otherwise the DWARF expression is ill-formed.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Could define this case as returning an implicit location description as
if the <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit</span></code> operation is performed.</p>
</div>
</li>
</ul>
</li>
<li><p>If the current result kind specifies a value, then:</p>
<ul class="simple">
<li><p>If the top stack entry is a value, or can be converted to one (see
<a class="reference internal" href="#amdgpu-dwarf-memory-location-description-operations"><span class="std std-ref">Memory Location Description Operations</span></a>), then the result
is that, possibly converted, value. Any other entries on the stack are
discarded.</p></li>
<li><p>Otherwise the DWARF expression is ill-formed.</p></li>
</ul>
</li>
<li><p>If the current result kind is not specified, then:</p>
<ul>
<li><p>If the stack is empty, the result is a location description with one
undefined location description.</p>
<p><em>This rule is for backwards compatibility with DWARF Version 5 which has no
explicit operation to create an undefined location description, and uses an
empty operation expression for this purpose.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This rule is consistent with the rule above for when a location
description is requested. However, GDB appears to report this as an error
and no GDB tests appear to cause an empty stack for this case.</p>
</div>
</li>
<li><p>Otherwise, the top stack entry is returned. Any other entries on the stack
are discarded.</p></li>
</ul>
</li>
</ul>
<p>An operation expression is encoded as a byte block with some form of prefix that
specifies the byte count. It can be used:</p>
<ul class="simple">
<li><p>as the value of a debugging information entry attribute that is encoded using
class <code class="docutils literal notranslate"><span class="pre">exprloc</span></code> (see DWARF Version 5 section 7.5.5),</p></li>
<li><p>as the operand to certain operation expression operations,</p></li>
<li><p>as the operand to certain call frame information operations (see
<a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>),</p></li>
<li><p>and in location list entries (see
<a class="reference internal" href="#amdgpu-dwarf-location-list-expressions"><span class="std std-ref">DWARF Location List Expressions</span></a>).</p></li>
</ul>
<div class="section" id="stack-operations">
<span id="amdgpu-dwarf-stack-operations"></span><h6><a class="toc-backref" href="#id14">Stack Operations</a><a class="headerlink" href="#stack-operations" title="Permalink to this headline">¶</a></h6>
<p>The following operations manipulate the DWARF stack. Operations that index the
stack assume that the top of the stack (most recently added entry) has index 0.
They allow the stack entries to be either a value or location description.</p>
<p>If any stack entry accessed by a stack operation is an incomplete composite
location description (see
<a class="reference internal" href="#amdgpu-dwarf-composite-location-description-operations"><span class="std std-ref">Composite Location Description Operations</span></a>), then the DWARF
expression is ill-formed.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>These operations now support stack entries that are values and location
descriptions.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If it is desired to also make them work with incomplete composite location
descriptions, then would need to define that the composite location storage
specified by the incomplete composite location description is also replicated
when a copy is pushed. This ensures that each copy of the incomplete composite
location description can update the composite location storage they specify
independently.</p>
</div>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_dup</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_dup</span></code> duplicates the stack entry at the top of the stack.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_drop</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_drop</span></code> pops the stack entry at the top of the stack and discards it.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_pick</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_pick</span></code> has a single unsigned 1-byte operand that represents an index
I. A copy of the stack entry with index I is pushed onto the stack.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_over</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_over</span></code> pushes a copy of the entry with index 1.</p>
<p><em>This is equivalent to a ``DW_OP_pick 1`` operation.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_swap</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_swap</span></code> swaps the top two stack entries. The entry at the top of the
stack becomes the second stack entry, and the second stack entry becomes the
top of the stack.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_rot</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_rot</span></code> rotates the first three stack entries. The entry at the top of
the stack becomes the third stack entry, the second entry becomes the top of
the stack, and the third entry becomes the second entry.</p>
</li>
</ol>
</div>
<div class="section" id="control-flow-operations">
<span id="amdgpu-dwarf-control-flow-operations"></span><h6><a class="toc-backref" href="#id15">Control Flow Operations</a><a class="headerlink" href="#control-flow-operations" title="Permalink to this headline">¶</a></h6>
<p>The following operations provide simple control of the flow of a DWARF operation
expression.</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_nop</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_nop</span></code> is a place holder. It has no effect on the DWARF stack
entries.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_le</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_ge</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_eq</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_lt</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_gt</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_OP_ne</span></code></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The same as in DWARF Version 5 section 2.5.1.5.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_skip</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_skip</span></code> is an unconditional branch. Its single operand is a 2-byte
signed integer constant. The 2-byte constant is the number of bytes of the
DWARF expression to skip forward or backward from the current operation,
beginning after the 2-byte constant.</p>
<p>If the updated position is at one past the end of the last operation, then
the operation expression evaluation is complete.</p>
<p>Otherwise, the DWARF expression is ill-formed if the updated operation
position is not in the range of the first to last operation inclusive, or
not at the start of an operation.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_bra</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_bra</span></code> is a conditional branch. Its single operand is a 2-byte signed
integer constant. This operation pops the top of stack. If the value popped
is not the constant 0, the 2-byte constant operand is the number of bytes of
the DWARF operation expression to skip forward or backward from the current
operation, beginning after the 2-byte constant.</p>
<p>If the updated position is at one past the end of the last operation, then
the operation expression evaluation is complete.</p>
<p>Otherwise, the DWARF expression is ill-formed if the updated operation
position is not in the range of the first to last operation inclusive, or
not at the start of an operation.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_call2,</span> <span class="pre">DW_OP_call4,</span> <span class="pre">DW_OP_call_ref</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_call2</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_call4</span></code>, and <code class="docutils literal notranslate"><span class="pre">DW_OP_call_ref</span></code> perform DWARF
procedure calls during evaluation of a DWARF expression.</p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_call2</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_OP_call4</span></code>, have one operand that is, respectively,
a 2-byte or 4-byte unsigned offset DR that represents the byte offset of a
debugging information entry D relative to the beginning of the current
compilation unit.</p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_call_ref</span></code> has one operand that is a 4-byte unsigned value in the
32-bit DWARF format, or an 8-byte unsigned value in the 64-bit DWARF format,
that represents the byte offset DR of a debugging information entry D
relative to the beginning of the <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code> section that contains the
current compilation unit. D may not be in the current compilation unit.</p>
<p><em>Operand interpretation of</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_call2</span></code><em>,</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_call4</span></code><em>, and</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_call_ref</span></code> <em>is exactly like that for</em> <code class="docutils literal notranslate"><span class="pre">DW_FORM_ref2</span></code><em>,
``DW_FORM_ref4``*, and</em> <code class="docutils literal notranslate"><span class="pre">DW_FORM_ref_addr</span></code><em>, respectively.</em></p>
<p>The call operation is evaluated by:</p>
<ul>
<li><p>If D has a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute that is encoded as a <code class="docutils literal notranslate"><span class="pre">exprloc</span></code>
that specifies an operation expression E, then execution of the current
operation expression continues from the first operation of E. Execution
continues until one past the last operation of E is reached, at which
point execution continues with the operation following the call operation.
The operations of E are evaluated with the same current context, except
current compilation unit is the one that contains D and the stack is the
same as that being used by the call operation. After the call operation
has been evaluated, the stack is therefore as it is left by the evaluation
of the operations of E. Since E is evaluated on the same stack as the call
operation, E can use, and/or remove entries already on the stack, and can
add new entries to the stack.</p>
<p><em>Values on the stack at the time of the call may be used as parameters by
the called expression and values left on the stack by the called expression
may be used as return values by prior agreement between the calling and
called expressions.</em></p>
</li>
<li><p>If D has a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute that is encoded as a <code class="docutils literal notranslate"><span class="pre">loclist</span></code> or
<code class="docutils literal notranslate"><span class="pre">loclistsptr</span></code>, then the specified location list expression E is
evaluated. The evaluation of E uses the current context, except the result
kind is a location description, the compilation unit is the one that
contains D, and the initial stack is empty. The location description
result is pushed on the stack.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This rule avoids having to define how to execute a matched location list
entry operation expression on the same stack as the call when there are
multiple matches. But it allows the call to obtain the location
description for a variable or formal parameter which may use a location
list expression.</p>
<p>An alternative is to treat the case when D has a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code>
attribute that is encoded as a <code class="docutils literal notranslate"><span class="pre">loclist</span></code> or <code class="docutils literal notranslate"><span class="pre">loclistsptr</span></code>, and the
specified location list expression E’ matches a single location list
entry with operation expression E, the same as the <code class="docutils literal notranslate"><span class="pre">exprloc</span></code> case and
evaluate on the same stack.</p>
<p>But this is not attractive as if the attribute is for a variable that
happens to end with a non-singleton stack, it will not simply put a
location description on the stack. Presumably the intent of using
<code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> on a variable or formal parameter debugger information
entry is to push just one location description on the stack. That
location description may have more than one single location description.</p>
<p>The previous rule for <code class="docutils literal notranslate"><span class="pre">exprloc</span></code> also has the same problem as normally
a variable or formal parameter location expression may leave multiple
entries on the stack and only return the top entry.</p>
<p>GDB implements <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> by always executing E on the same stack.
If the location list has multiple matching entries, it simply picks the
first one and ignores the rest. This seems fundamentally at odds with
the desire to supporting multiple places for variables.</p>
<p>So, it feels like <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> should both support pushing a location
description on the stack for a variable or formal parameter, and also
support being able to execute an operation expression on the same stack.
Being able to specify a different operation expression for different
program locations seems a desirable feature to retain.</p>
<p>A solution to that is to have a distinct <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_proc</span></code> attribute
for the <code class="docutils literal notranslate"><span class="pre">DW_TAG_dwarf_procedure</span></code> debugging information entry. Then the
<code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute expression is always executed separately
and pushes a location description (that may have multiple single
location descriptions), and the <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_proc</span></code> attribute expression
is always executed on the same stack and can leave anything on the
stack.</p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_proc</span></code> attribute could have the new classes
<code class="docutils literal notranslate"><span class="pre">exprproc</span></code>, <code class="docutils literal notranslate"><span class="pre">loclistproc</span></code>, and <code class="docutils literal notranslate"><span class="pre">loclistsptrproc</span></code> to indicate that
the expression is executed on the same stack. <code class="docutils literal notranslate"><span class="pre">exprproc</span></code> is the same
encoding as <code class="docutils literal notranslate"><span class="pre">exprloc</span></code>. <code class="docutils literal notranslate"><span class="pre">loclistproc</span></code> and <code class="docutils literal notranslate"><span class="pre">loclistsptrproc</span></code> are the
same encoding as their non-<code class="docutils literal notranslate"><span class="pre">proc</span></code> counterparts, except the DWARF is
ill-formed if the location list does not match exactly one location list
entry and a default entry is required. These forms indicate explicitly
that the matched single operation expression must be executed on the
same stack. This is better than ad hoc special rules for <code class="docutils literal notranslate"><span class="pre">loclistproc</span></code>
and <code class="docutils literal notranslate"><span class="pre">loclistsptrproc</span></code> which are currently clearly defined to always
return a location description. The producer then explicitly indicates
the intent through the attribute classes.</p>
<p>Such a change would be a breaking change for how GDB implements
<code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code>. However, are the breaking cases actually occurring in
practice? GDB could implement the current approach for DWARF Version 5,
and the new semantics for DWARF Version 6 which has been done for some
other features.</p>
<p>Another option is to limit the execution to be on the same stack only to
the evaluation of an expression E that is the value of a
<code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute of a <code class="docutils literal notranslate"><span class="pre">DW_TAG_dwarf_procedure</span></code> debugging
information entry. The DWARF would be ill-formed if E is a location list
expression that does not match exactly one location list entry. In all
other cases the evaluation of an expression E that is the value of a
<code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute would evaluate E with the current context,
except the result kind is a location description, the compilation unit
is the one that contains D, and the initial stack is empty. The location
description result is pushed on the stack.</p>
</div>
</li>
<li><p>If D has a <code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code> attribute with a value V, then it is as
if a <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_value</span> <span class="pre">V</span></code> operation was executed.</p>
<p><em>This allows a call operation to be used to compute the location
description for any variable or formal parameter regardless of whether the
producer has optimized it to a constant. This is consistent with the
``DW_OP_implicit_pointer`` operation.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Alternatively, could deprecate using <code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code> for
<code class="docutils literal notranslate"><span class="pre">DW_TAG_variable</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_TAG_formal_parameter</span></code> debugger information
entries that are constants and instead use <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> with an
operation expression that results in a location description with one
implicit location description. Then this rule would not be required.</p>
</div>
</li>
<li><p>Otherwise, there is no effect and no changes are made to the stack.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>In DWARF Version 5, if D does not have a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> then
<code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> is defined to have no effect. It is unclear that this is
the right definition as a producer should be able to rely on using
<code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> to get a location description for any non-<code class="docutils literal notranslate"><span class="pre">DW_TAG_dwarf_procedure</span></code> debugging information entries. Also, the
producer should not be creating DWARF with <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> to a
<code class="docutils literal notranslate"><span class="pre">DW_TAG_dwarf_procedure</span></code> that does not have a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code>
attribute. So, should this case be defined as an ill-formed DWARF
expression?</p>
</div>
</li>
</ul>
<p><em>The</em> <code class="docutils literal notranslate"><span class="pre">DW_TAG_dwarf_procedure</span></code> <em>debugging information entry can be used to
define DWARF procedures that can be called.</em></p>
</li>
</ol>
</div>
<div class="section" id="value-operations">
<span id="amdgpu-dwarf-value-operations"></span><h6><a class="toc-backref" href="#id16">Value Operations</a><a class="headerlink" href="#value-operations" title="Permalink to this headline">¶</a></h6>
<p>This section describes the operations that push values on the stack.</p>
<p>Each value stack entry has a type and a literal value and can represent a
literal value of any supported base type of the target architecture. The base
type specifies the size, encoding, and endianity of the literal value.</p>
<p>The base type of value stack entries can be the distinguished generic type.</p>
<div class="section" id="literal-operations">
<span id="amdgpu-dwarf-literal-operations"></span><h7><a class="toc-backref" href="#id17">Literal Operations</a><a class="headerlink" href="#literal-operations" title="Permalink to this headline">¶</a></h7>
<p>The following operations all push a literal value onto the DWARF stack.</p>
<p>Operations other than <code class="docutils literal notranslate"><span class="pre">DW_OP_const_type</span></code> push a value V with the generic type.
If V is larger than the generic type, then V is truncated to the generic type
size and the low-order bits used.</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_lit0</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_lit1</span></code>, …, <code class="docutils literal notranslate"><span class="pre">DW_OP_lit31</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_lit<N></span></code> operations encode an unsigned literal value N from 0
through 31, inclusive. They push the value N with the generic type.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_const1u</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_const2u</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_const4u</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_const8u</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_const<N>u</span></code> operations have a single operand that is a 1, 2, 4, or
8-byte unsigned integer constant U, respectively. They push the value U with
the generic type.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_const1s</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_const2s</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_const4s</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_const8s</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_const<N>s</span></code> operations have a single operand that is a 1, 2, 4, or
8-byte signed integer constant S, respectively. They push the value S with
the generic type.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span></code> has a single unsigned LEB128 integer operand N. It pushes
the value N with the generic type.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_consts</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_consts</span></code> has a single signed LEB128 integer operand N. It pushes the
value N with the generic type.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_constx</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_constx</span></code> has a single unsigned LEB128 integer operand that
represents a zero-based index into the <code class="docutils literal notranslate"><span class="pre">.debug_addr</span></code> section relative to
the value of the <code class="docutils literal notranslate"><span class="pre">DW_AT_addr_base</span></code> attribute of the associated compilation
unit. The value N in the <code class="docutils literal notranslate"><span class="pre">.debug_addr</span></code> section has the size of the generic
type. It pushes the value N with the generic type.</p>
<p><em>The</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_constx</span></code> <em>operation is provided for constants that require
link-time relocation but should not be interpreted by the consumer as a
relocatable address (for example, offsets to thread-local storage).</em></p>
</li>
</ol>
<ol class="arabic" start="9">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_const_type</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_const_type</span></code> has three operands. The first is an unsigned LEB128
integer DR that represents the byte offset of a debugging information entry
D relative to the beginning of the current compilation unit, that provides
the type T of the constant value. The second is a 1-byte unsigned integral
constant S. The third is a block of bytes B, with a length equal to S.</p>
<p>TS is the bit size of the type T. The least significant TS bits of B are
interpreted as a value V of the type D. It pushes the value V with the type
D.</p>
<p>The DWARF is ill-formed if D is not a <code class="docutils literal notranslate"><span class="pre">DW_TAG_base_type</span></code> debugging
information entry in the current compilation unit, or if TS divided by 8
(the byte size) and rounded up to a whole number is not equal to S.</p>
<p><em>While the size of the byte block B can be inferred from the type D
definition, it is encoded explicitly into the operation so that the
operation can be parsed easily without reference to the</em> <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code>
<em>section.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_push_lane</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_push_lane</span></code> pushes the target architecture lane identifier of
the current lane as a value with the generic type.</p>
<p><em>For languages that are implemented using a SIMD or SIMT execution model,
this is the lane number that corresponds to the source language thread of
execution upon which the user is focused.</em></p>
</li>
</ol>
</div>
<div class="section" id="arithmetic-and-logical-operations">
<span id="amdgpu-dwarf-arithmetic-logical-operations"></span><h7><a class="toc-backref" href="#id18">Arithmetic and Logical Operations</a><a class="headerlink" href="#arithmetic-and-logical-operations" title="Permalink to this headline">¶</a></h7>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This section is the same as DWARF Version 5 section 2.5.1.4.</p>
</div>
</div>
<div class="section" id="type-conversion-operations">
<span id="amdgpu-dwarf-type-conversions-operations"></span><h7><a class="toc-backref" href="#id19">Type Conversion Operations</a><a class="headerlink" href="#type-conversion-operations" title="Permalink to this headline">¶</a></h7>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This section is the same as DWARF Version 5 section 2.5.1.6.</p>
</div>
</div>
<div class="section" id="special-value-operations">
<span id="amdgpu-dwarf-general-operations"></span><h7><a class="toc-backref" href="#id20">Special Value Operations</a><a class="headerlink" href="#special-value-operations" title="Permalink to this headline">¶</a></h7>
<p>There are these special value operations currently defined:</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_regval_type</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_regval_type</span></code> has two operands. The first is an unsigned LEB128
integer that represents a register number R. The second is an unsigned
LEB128 integer DR that represents the byte offset of a debugging information
entry D relative to the beginning of the current compilation unit, that
provides the type T of the register value.</p>
<p>The operation is equivalent to performing <code class="docutils literal notranslate"><span class="pre">DW_OP_regx</span> <span class="pre">R;</span> <span class="pre">DW_OP_deref_type</span>
<span class="pre">DR</span></code>.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Should DWARF allow the type T to be a larger size than the size of the
register R? Restricting a larger bit size avoids any issue of conversion
as the, possibly truncated, bit contents of the register is simply
interpreted as a value of T. If a conversion is wanted it can be done
explicitly using a <code class="docutils literal notranslate"><span class="pre">DW_OP_convert</span></code> operation.</p>
<p>GDB has a per register hook that allows a target specific conversion on a
register by register basis. It defaults to truncation of bigger registers.
Removing use of the target hook does not cause any test failures in common
architectures. If the compiler for a target architecture did want some
form of conversion, including a larger result type, it could always
explicitly used the <code class="docutils literal notranslate"><span class="pre">DW_OP_convert</span></code> operation.</p>
<p>If T is a larger type than the register size, then the default GDB
register hook reads bytes from the next register (or reads out of bounds
for the last register!). Removing use of the target hook does not cause
any test failures in common architectures (except an illegal hand written
assembly test). If a target architecture requires this behavior, these
extensions allow a composite location description to be used to combine
multiple registers.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_deref</span></code></p>
<p>S is the bit size of the generic type divided by 8 (the byte size) and
rounded up to a whole number. DR is the offset of a hypothetical debug
information entry D in the current compilation unit for a base type of the
generic type.</p>
<p>The operation is equivalent to performing <code class="docutils literal notranslate"><span class="pre">DW_OP_deref_type</span> <span class="pre">S,</span> <span class="pre">DR</span></code>.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_deref_size</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_deref_size</span></code> has a single 1-byte unsigned integral constant that
represents a byte result size S.</p>
<p>TS is the smaller of the generic type bit size and S scaled by 8 (the byte
size). If TS is smaller than the generic type bit size then T is an unsigned
integral type of bit size TS, otherwise T is the generic type. DR is the
offset of a hypothetical debug information entry D in the current
compilation unit for a base type T.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Truncating the value when S is larger than the generic type matches what
GDB does. This allows the generic type size to not be an integral byte
size. It does allow S to be arbitrarily large. Should S be restricted to
the size of the generic type rounded up to a multiple of 8?</p>
</div>
<p>The operation is equivalent to performing <code class="docutils literal notranslate"><span class="pre">DW_OP_deref_type</span> <span class="pre">S,</span> <span class="pre">DR</span></code>, except
if T is not the generic type, the value V pushed is zero-extended to the
generic type bit size and its type changed to the generic type.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_deref_type</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_deref_type</span></code> has two operands. The first is a 1-byte unsigned
integral constant S. The second is an unsigned LEB128 integer DR that
represents the byte offset of a debugging information entry D relative to
the beginning of the current compilation unit, that provides the type T of
the result value.</p>
<p>TS is the bit size of the type T.</p>
<p><em>While the size of the pushed value V can be inferred from the type T, it is
encoded explicitly as the operand S so that the operation can be parsed
easily without reference to the</em> <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code> <em>section.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>It is unclear why the operand S is needed. Unlike <code class="docutils literal notranslate"><span class="pre">DW_OP_const_type</span></code>,
the size is not needed for parsing. Any evaluation needs to get the base
type T to push with the value to know its encoding and bit size.</p>
</div>
<p>It pops one stack entry that must be a location description L.</p>
<p>A value V of TS bits is retrieved from the location storage LS specified by
one of the single location descriptions SL of L.</p>
<p><em>If L, or the location description of any composite location description
part that is a subcomponent of L, has more than one single location
description, then any one of them can be selected as they are required to
all have the same value. For any single location description SL, bits are
retrieved from the associated storage location starting at the bit offset
specified by SL. For a composite location description, the retrieved bits
are the concatenation of the N bits from each composite location part PL,
where N is limited to the size of PL.</em></p>
<p>V is pushed on the stack with the type T.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This definition makes it an evaluation error if L is a register location
description that has less than TS bits remaining in the register storage.
Particularly since these extensions extend location descriptions to have
a bit offset, it would be odd to define this as performing sign extension
based on the type, or be target architecture dependent, as the number of
remaining bits could be any number. This matches the GDB implementation
for <code class="docutils literal notranslate"><span class="pre">DW_OP_deref_type</span></code>.</p>
<p>These extensions define <code class="docutils literal notranslate"><span class="pre">DW_OP_*breg*</span></code> in terms of
<code class="docutils literal notranslate"><span class="pre">DW_OP_regval_type</span></code>. <code class="docutils literal notranslate"><span class="pre">DW_OP_regval_type</span></code> is defined in terms of
<code class="docutils literal notranslate"><span class="pre">DW_OP_regx</span></code>, which uses a 0 bit offset, and <code class="docutils literal notranslate"><span class="pre">DW_OP_deref_type</span></code>.
Therefore, it requires the register size to be greater or equal to the
address size of the address space. This matches the GDB implementation for
<code class="docutils literal notranslate"><span class="pre">DW_OP_*breg*</span></code>.</p>
</div>
<p>The DWARF is ill-formed if D is not in the current compilation unit, D is
not a <code class="docutils literal notranslate"><span class="pre">DW_TAG_base_type</span></code> debugging information entry, or if TS divided by
8 (the byte size) and rounded up to a whole number is not equal to S.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This definition allows the base type to be a bit size since there seems no
reason to restrict it.</p>
</div>
<p>It is an evaluation error if any bit of the value is retrieved from the
undefined location storage or the offset of any bit exceeds the size of the
location storage LS specified by any single location description SL of L.</p>
<p>See <a class="reference internal" href="#amdgpu-dwarf-implicit-location-descriptions"><span class="std std-ref">Implicit Location Description Operations</span></a> for special rules
concerning implicit location descriptions created by the
<code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_implicit_aspace_pointer</span></code>
operations.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_xderef</span></code> <em>Deprecated</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_xderef</span></code> pops two stack entries. The first must be an integral type
value that represents an address A. The second must be an integral type
value that represents a target architecture specific address space
identifier AS.</p>
<p>The operation is equivalent to performing <code class="docutils literal notranslate"><span class="pre">DW_OP_swap;</span>
<span class="pre">DW_OP_LLVM_form_aspace_address;</span> <span class="pre">DW_OP_deref</span></code>. The value V retrieved is left
on the stack with the generic type.</p>
<p><em>This operation is deprecated as the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code>
<em>operation can be used and provides greater expressiveness.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_xderef_size</span></code> <em>Deprecated</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_xderef_size</span></code> has a single 1-byte unsigned integral constant that
represents a byte result size S.</p>
<p>It pops two stack entries. The first must be an integral type value that
represents an address A. The second must be an integral type value that
represents a target architecture specific address space identifier AS.</p>
<p>The operation is equivalent to performing <code class="docutils literal notranslate"><span class="pre">DW_OP_swap;</span>
<span class="pre">DW_OP_LLVM_form_aspace_address;</span> <span class="pre">DW_OP_deref_size</span> <span class="pre">S</span></code>. The zero-extended
value V retrieved is left on the stack with the generic type.</p>
<p><em>This operation is deprecated as the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code>
<em>operation can be used and provides greater expressiveness.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_xderef_type</span></code> <em>Deprecated</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_xderef_type</span></code> has two operands. The first is a 1-byte unsigned
integral constant S. The second operand is an unsigned LEB128 integer DR
that represents the byte offset of a debugging information entry D relative
to the beginning of the current compilation unit, that provides the type T
of the result value.</p>
<p>It pops two stack entries. The first must be an integral type value that
represents an address A. The second must be an integral type value that
represents a target architecture specific address space identifier AS.</p>
<p>The operation is equivalent to performing <code class="docutils literal notranslate"><span class="pre">DW_OP_swap;</span>
<span class="pre">DW_OP_LLVM_form_aspace_address;</span> <span class="pre">DW_OP_deref_type</span> <span class="pre">S</span> <span class="pre">R</span></code>. The value V
retrieved is left on the stack with the type D.</p>
<p><em>This operation is deprecated as the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code>
<em>operation can be used and provides greater expressiveness.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_entry_value</span></code> <em>Deprecated</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_entry_value</span></code> pushes the value of an expression that is evaluated in
the context of the calling frame.</p>
<p><em>It may be used to determine the value of arguments on entry to the current
call frame provided they are not clobbered.</em></p>
<p>It has two operands. The first is an unsigned LEB128 integer S. The second
is a block of bytes, with a length equal S, interpreted as a DWARF
operation expression E.</p>
<p>E is evaluated with the current context, except the result kind is
unspecified, the call frame is the one that called the current frame, the
program location is the call site in the calling frame, the object is
unspecified, and the initial stack is empty. The calling frame information
is obtained by virtually unwinding the current call frame using the call
frame information (see <a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>).</p>
<p>If the result of E is a location description L (see
<a class="reference internal" href="#amdgpu-dwarf-register-location-descriptions"><span class="std std-ref">Register Location Description Operations</span></a>), and the last operation
executed by E is a <code class="docutils literal notranslate"><span class="pre">DW_OP_reg*</span></code> for register R with a target architecture
specific base type of T, then the contents of the register are retrieved as
if a <code class="docutils literal notranslate"><span class="pre">DW_OP_deref_type</span> <span class="pre">DR</span></code> operation was performed where DR is the offset
of a hypothetical debug information entry in the current compilation unit
for T. The resulting value V s pushed on the stack.</p>
<p><em>Using</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_reg*</span></code> <em>provides a more compact form for the case where the
value was in a register on entry to the subprogram.</em></p>
<p>If the result of E is a value V, then V is pushed on the stack.</p>
<p>Otherwise, the DWARF expression is ill-formed.</p>
<p><em>The</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_entry_value</span></code> <em>operation is deprecated as its main usage is
provided by other means. DWARF Version 5 added the</em>
<code class="docutils literal notranslate"><span class="pre">DW_TAG_call_site_parameter</span></code> <em>debugger information entry for call sites
that has</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_call_value</span></code><em>,</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_location</span></code><em>, and</em>
<code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_value</span></code> <em>attributes that provide DWARF expressions to
compute actual parameter values at the time of the call, and requires the
producer to ensure the expressions are valid to evaluate even when virtually
unwound. The</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_call_frame_entry_reg</span></code> <em>operation provides access
to registers in the virtually unwound calling frame.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>GDB only implements <code class="docutils literal notranslate"><span class="pre">DW_OP_entry_value</span></code> when E is exactly
<code class="docutils literal notranslate"><span class="pre">DW_OP_reg*</span></code> or <code class="docutils literal notranslate"><span class="pre">DW_OP_breg*;</span> <span class="pre">DW_OP_deref*</span></code>.</p>
</div>
</li>
</ol>
</div>
</div>
<div class="section" id="location-description-operations">
<span id="amdgpu-dwarf-location-description-operations"></span><h6><a class="toc-backref" href="#id21">Location Description Operations</a><a class="headerlink" href="#location-description-operations" title="Permalink to this headline">¶</a></h6>
<p>This section describes the operations that push location descriptions on the
stack.</p>
<div class="section" id="general-location-description-operations">
<h7><a class="toc-backref" href="#id22">General Location Description Operations</a><a class="headerlink" href="#general-location-description-operations" title="Permalink to this headline">¶</a></h7>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span></code> pops two stack entries. The first must be an integral
type value that represents a byte displacement B. The second must be a
location description L.</p>
<p>It adds the value of B scaled by 8 (the byte size) to the bit offset of each
single location description SL of L, and pushes the updated L.</p>
<p>It is an evaluation error if the updated bit offset of any SL is less than 0
or greater than or equal to the size of the location storage specified by
SL.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset_uconst</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset_uconst</span></code> has a single unsigned LEB128 integer operand
that represents a byte displacement B.</p>
<p>The operation is equivalent to performing <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">B;</span>
<span class="pre">DW_OP_LLVM_offset</span></code>.</p>
<p><em>This operation is supplied specifically to be able to encode more field
displacements in two bytes than can be done with</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_lit*;</span>
<span class="pre">DW_OP_LLVM_offset</span></code><em>.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Should this be named <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset_uconst</span></code> to match
<code class="docutils literal notranslate"><span class="pre">DW_OP_plus_uconst</span></code>, or <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset_constu</span></code> to match
<code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span></code>?</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_bit_offset</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_bit_offset</span></code> pops two stack entries. The first must be an
integral type value that represents a bit displacement B. The second must be
a location description L.</p>
<p>It adds the value of B to the bit offset of each single location description
SL of L, and pushes the updated L.</p>
<p>It is an evaluation error if the updated bit offset of any SL is less than 0
or greater than or equal to the size of the location storage specified by
SL.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> pushes the location description L of the
current object.</p>
<p><em>This object may correspond to an independent variable that is part of a
user presented expression that is being evaluated. The object location
description may be determined from the variable’s own debugging information
entry or it may be a component of an array, structure, or class whose
address has been dynamically determined by an earlier step during user
expression evaluation.</em></p>
<p><em>This operation provides explicit functionality (especially for arrays
involving descriptions) that is analogous to the implicit push of the base
location description of a structure prior to evaluation of a
``DW_AT_data_member_location`` to access a data member of a structure.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This operation could be removed and the object location description
specified as the initial stack as for <code class="docutils literal notranslate"><span class="pre">DW_AT_data_member_location</span></code>.</p>
<p>The only attribute that specifies a current object is
<code class="docutils literal notranslate"><span class="pre">DW_AT_data_location</span></code> so the non-normative text seems to overstate how
this is being used. Or are there other attributes that need to state they
pass an object?</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_call_frame_entry_reg</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_call_frame_entry_reg</span></code> has a single unsigned LEB128 integer
operand that represents a target architecture register number R.</p>
<p>It pushes a location description L that holds the value of register R on
entry to the current subprogram as defined by the call frame information
(see <a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>).</p>
<p><em>If there is no call frame information defined, then the default rules for
the target architecture are used. If the register rule is</em> undefined<em>, then
the undefined location description is pushed. If the register rule is</em> same
value<em>, then a register location description for R is pushed.</em></p>
</li>
</ol>
</div>
<div class="section" id="undefined-location-description-operations">
<span id="amdgpu-dwarf-undefined-location-description-operations"></span><h7><a class="toc-backref" href="#id23">Undefined Location Description Operations</a><a class="headerlink" href="#undefined-location-description-operations" title="Permalink to this headline">¶</a></h7>
<p><em>The undefined location storage represents a piece or all of an object that is
present in the source but not in the object code (perhaps due to optimization).
Neither reading nor writing to the undefined location storage is meaningful.</em></p>
<p>An undefined location description specifies the undefined location storage.
There is no concept of the size of the undefined location storage, nor of a bit
offset for an undefined location description. The <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_*offset</span></code>
operations leave an undefined location description unchanged. The
<code class="docutils literal notranslate"><span class="pre">DW_OP_*piece</span></code> operations can explicitly or implicitly specify an undefined
location description, allowing any size and offset to be specified, and results
in a part with all undefined bits.</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_undefined</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_undefined</span></code> pushes a location description L that comprises one
undefined location description SL.</p>
</li>
</ol>
</div>
<div class="section" id="memory-location-description-operations">
<span id="amdgpu-dwarf-memory-location-description-operations"></span><h7><a class="toc-backref" href="#id24">Memory Location Description Operations</a><a class="headerlink" href="#memory-location-description-operations" title="Permalink to this headline">¶</a></h7>
<p>Each of the target architecture specific address spaces has a corresponding
memory location storage that denotes the linear addressable memory of that
address space. The size of each memory location storage corresponds to the range
of the addresses in the corresponding address space.</p>
<p><em>It is target architecture defined how address space location storage maps to
target architecture physical memory. For example, they may be independent
memory, or more than one location storage may alias the same physical memory
possibly at different offsets and with different interleaving. The mapping may
also be dictated by the source language address classes.</em></p>
<p>A memory location description specifies a memory location storage. The bit
offset corresponds to a bit position within a byte of the memory. Bits accessed
using a memory location description, access the corresponding target
architecture memory starting at the bit position within the byte specified by
the bit offset.</p>
<p>A memory location description that has a bit offset that is a multiple of 8 (the
byte size) is defined to be a byte address memory location description. It has a
memory byte address A that is equal to the bit offset divided by 8.</p>
<p>A memory location description that does not have a bit offset that is a multiple
of 8 (the byte size) is defined to be a bit field memory location description.
It has a bit position B equal to the bit offset modulo 8, and a memory byte
address A equal to the bit offset minus B that is then divided by 8.</p>
<p>The address space AS of a memory location description is defined to be the
address space that corresponds to the memory location storage associated with
the memory location description.</p>
<p>A location description that is comprised of one byte address memory location
description SL is defined to be a memory byte address location description. It
has a byte address equal to A and an address space equal to AS of the
corresponding SL.</p>
<p><code class="docutils literal notranslate"><span class="pre">DW_ASPACE_none</span></code> is defined as the target architecture default address space.</p>
<p>If a stack entry is required to be a location description, but it is a value V
with the generic type, then it is implicitly converted to a location description
L with one memory location description SL. SL specifies the memory location
storage that corresponds to the target architecture default address space with a
bit offset equal to V scaled by 8 (the byte size).</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If it is wanted to allow any integral type value to be implicitly converted to
a memory location description in the target architecture default address
space:</p>
<blockquote>
<div><p>If a stack entry is required to be a location description, but is a value V
with an integral type, then it is implicitly converted to a location
description L with a one memory location description SL. If the type size of
V is less than the generic type size, then the value V is zero extended to
the size of the generic type. The least significant generic type size bits
are treated as a twos-complement unsigned value to be used as an address A.
SL specifies memory location storage corresponding to the target
architecture default address space with a bit offset equal to A scaled by 8
(the byte size).</p>
</div></blockquote>
<p>The implicit conversion could also be defined as target architecture specific.
For example, GDB checks if V is an integral type. If it is not it gives an
error. Otherwise, GDB zero-extends V to 64 bits. If the GDB target defines a
hook function, then it is called. The target specific hook function can modify
the 64-bit value, possibly sign extending based on the original value type.
Finally, GDB treats the 64-bit value V as a memory location address.</p>
</div>
<p>If a stack entry is required to be a location description, but it is an implicit
pointer value IPV with the target architecture default address space, then it is
implicitly converted to a location description with one single location
description specified by IPV. See
<a class="reference internal" href="#amdgpu-dwarf-implicit-location-descriptions"><span class="std std-ref">Implicit Location Description Operations</span></a>.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Is this rule required for DWARF Version 5 backwards compatibility? If not, it
can be eliminated, and the producer can use
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code>.</p>
</div>
<p>If a stack entry is required to be a value, but it is a location description L
with one memory location description SL in the target architecture default
address space with a bit offset B that is a multiple of 8, then it is implicitly
converted to a value equal to B divided by 8 (the byte size) with the generic
type.</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_addr</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_addr</span></code> has a single byte constant value operand, which has the size
of the generic type, that represents an address A.</p>
<p>It pushes a location description L with one memory location description SL
on the stack. SL specifies the memory location storage corresponding to the
target architecture default address space with a bit offset equal to A
scaled by 8 (the byte size).</p>
<p><em>If the DWARF is part of a code object, then A may need to be relocated. For
example, in the ELF code object format, A must be adjusted by the difference
between the ELF segment virtual address and the virtual address at which the
segment is loaded.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_addrx</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_addrx</span></code> has a single unsigned LEB128 integer operand that represents
a zero-based index into the <code class="docutils literal notranslate"><span class="pre">.debug_addr</span></code> section relative to the value of
the <code class="docutils literal notranslate"><span class="pre">DW_AT_addr_base</span></code> attribute of the associated compilation unit. The
address value A in the <code class="docutils literal notranslate"><span class="pre">.debug_addr</span></code> section has the size of the generic
type.</p>
<p>It pushes a location description L with one memory location description SL
on the stack. SL specifies the memory location storage corresponding to the
target architecture default address space with a bit offset equal to A
scaled by 8 (the byte size).</p>
<p><em>If the DWARF is part of a code object, then A may need to be relocated. For
example, in the ELF code object format, A must be adjusted by the difference
between the ELF segment virtual address and the virtual address at which the
segment is loaded.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code> pops top two stack entries. The first
must be an integral type value that represents a target architecture
specific address space identifier AS. The second must be an integral type
value that represents an address A.</p>
<p>The address size S is defined as the address bit size of the target
architecture specific address space that corresponds to AS.</p>
<p>A is adjusted to S bits by zero extending if necessary, and then treating the
least significant S bits as a twos-complement unsigned value A’.</p>
<p>It pushes a location description L with one memory location description SL
on the stack. SL specifies the memory location storage LS that corresponds
to AS with a bit offset equal to A’ scaled by 8 (the byte size).</p>
<p>If AS is an address space that is specific to context elements, then LS
corresponds to the location storage associated with the current context.</p>
<p><em>For example, if AS is for per thread storage then LS is the location
storage for the current thread. For languages that are implemented using a
SIMD or SIMT execution model, then if AS is for per lane storage then LS is
the location storage for the current lane of the current thread. Therefore,
if L is accessed by an operation, the location storage selected when the
location description was created is accessed, and not the location storage
associated with the current context of the access operation.</em></p>
<p>The DWARF expression is ill-formed if AS is not one of the values defined by
the target architecture specific <code class="docutils literal notranslate"><span class="pre">DW_ASPACE_*</span></code> values.</p>
<p>See <a class="reference internal" href="#amdgpu-dwarf-implicit-location-descriptions"><span class="std std-ref">Implicit Location Description Operations</span></a> for special rules
concerning implicit pointer values produced by dereferencing implicit
location descriptions created by the <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> and
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_implicit_aspace_pointer</span></code> operations.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_form_tls_address</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_form_tls_address</span></code> pops one stack entry that must be an integral
type value and treats it as a thread-local storage address TA.</p>
<p>It pushes a location description L with one memory location description SL
on the stack. SL is the target architecture specific memory location
description that corresponds to the thread-local storage address TA.</p>
<p>The meaning of the thread-local storage address TA is defined by the
run-time environment. If the run-time environment supports multiple
thread-local storage blocks for a single thread, then the block
corresponding to the executable or shared library containing this DWARF
expression is used.</p>
<p><em>Some implementations of C, C++, Fortran, and other languages support a
thread-local storage class. Variables with this storage class have distinct
values and addresses in distinct threads, much as automatic variables have
distinct values and addresses in each subprogram invocation. Typically,
there is a single block of storage containing all thread-local variables
declared in the main executable, and a separate block for the variables
declared in each shared library. Each thread-local variable can then be
accessed in its block using an identifier. This identifier is typically a
byte offset into the block and pushed onto the DWARF stack by one of the</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_const*</span></code> <em>operations prior to the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_form_tls_address</span></code>
<em>operation. Computing the address of the appropriate block can be complex
(in some cases, the compiler emits a function call to do it), and difficult
to describe using ordinary DWARF location descriptions. Instead of forcing
complex thread-local storage calculations into the DWARF expressions, the</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_form_tls_address</span></code> <em>allows the consumer to perform the computation
based on the target architecture specific run-time environment.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_call_frame_cfa</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_call_frame_cfa</span></code> pushes the location description L of the Canonical
Frame Address (CFA) of the current subprogram, obtained from the call frame
information on the stack. See <a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>.</p>
<p><em>Although the value of the</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_frame_base</span></code> <em>attribute of the debugger
information entry corresponding to the current subprogram can be computed
using a location list expression, in some cases this would require an
extensive location list because the values of the registers used in
computing the CFA change during a subprogram execution. If the call frame
information is present, then it already encodes such changes, and it is
space efficient to reference that using the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_call_frame_cfa</span></code>
<em>operation.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_fbreg</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_fbreg</span></code> has a single signed LEB128 integer operand that represents a
byte displacement B.</p>
<p>The location description L for the <em>frame base</em> of the current subprogram is
obtained from the <code class="docutils literal notranslate"><span class="pre">DW_AT_frame_base</span></code> attribute of the debugger information
entry corresponding to the current subprogram as described in
<a class="reference internal" href="#amdgpu-dwarf-debugging-information-entry-attributes"><span class="std std-ref">Debugging Information Entry Attributes</span></a>.</p>
<p>The location description L is updated as if the <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset_uconst</span>
<span class="pre">B</span></code> operation was applied. The updated L is pushed on the stack.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_breg0</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_breg1</span></code>, …, <code class="docutils literal notranslate"><span class="pre">DW_OP_breg31</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_OP_breg<N></span></code> operations encode the numbers of up to 32 registers,
numbered from 0 through 31, inclusive. The register number R corresponds to
the N in the operation name.</p>
<p>They have a single signed LEB128 integer operand that represents a byte
displacement B.</p>
<p>The address space identifier AS is defined as the one corresponding to the
target architecture specific default address space.</p>
<p>The address size S is defined as the address bit size of the target
architecture specific address space corresponding to AS.</p>
<p>The contents of the register specified by R are retrieved as if a
<code class="docutils literal notranslate"><span class="pre">DW_OP_regval_type</span> <span class="pre">R,</span> <span class="pre">DR</span></code> operation was performed where DR is the offset
of a hypothetical debug information entry in the current compilation unit
for an unsigned integral base type of size S bits. B is added and the least
significant S bits are treated as an unsigned value to be used as an address
A.</p>
<p>They push a location description L comprising one memory location
description LS on the stack. LS specifies the memory location storage that
corresponds to AS with a bit offset equal to A scaled by 8 (the byte size).</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_bregx</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_bregx</span></code> has two operands. The first is an unsigned LEB128 integer
that represents a register number R. The second is a signed LEB128
integer that represents a byte displacement B.</p>
<p>The action is the same as for <code class="docutils literal notranslate"><span class="pre">DW_OP_breg<N></span></code>, except that R is used as
the register number and B is used as the byte displacement.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_bregx</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_bregx</span></code> has two operands. The first is an unsigned
LEB128 integer that represents a register number R. The second is a signed
LEB128 integer that represents a byte displacement B. It pops one stack
entry that is required to be an integral type value that represents a target
architecture specific address space identifier AS.</p>
<p>The action is the same as for <code class="docutils literal notranslate"><span class="pre">DW_OP_breg<N></span></code>, except that R is used as
the register number, B is used as the byte displacement, and AS is used as
the address space identifier.</p>
<p>The DWARF expression is ill-formed if AS is not one of the values defined by
the target architecture specific <code class="docutils literal notranslate"><span class="pre">DW_ASPACE_*</span></code> values.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Could also consider adding <code class="docutils literal notranslate"><span class="pre">DW_OP_aspace_breg0,</span> <span class="pre">DW_OP_aspace_breg1,</span> <span class="pre">...,</span>
<span class="pre">DW_OP_aspace_bref31</span></code> which would save encoding size.</p>
</div>
</li>
</ol>
</div>
<div class="section" id="register-location-description-operations">
<span id="amdgpu-dwarf-register-location-descriptions"></span><h7><a class="toc-backref" href="#id25">Register Location Description Operations</a><a class="headerlink" href="#register-location-description-operations" title="Permalink to this headline">¶</a></h7>
<p>There is a register location storage that corresponds to each of the target
architecture registers. The size of each register location storage corresponds
to the size of the corresponding target architecture register.</p>
<p>A register location description specifies a register location storage. The bit
offset corresponds to a bit position within the register. Bits accessed using a
register location description access the corresponding target architecture
register starting at the specified bit offset.</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_reg0</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_reg1</span></code>, …, <code class="docutils literal notranslate"><span class="pre">DW_OP_reg31</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_reg<N></span></code> operations encode the numbers of up to 32 registers,
numbered from 0 through 31, inclusive. The target architecture register
number R corresponds to the N in the operation name.</p>
<p>The operation is equivalent to performing <code class="docutils literal notranslate"><span class="pre">DW_OP_regx</span> <span class="pre">R</span></code>.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_regx</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_regx</span></code> has a single unsigned LEB128 integer operand that represents
a target architecture register number R.</p>
<p>If the current call frame is the top call frame, it pushes a location
description L that specifies one register location description SL on the
stack. SL specifies the register location storage that corresponds to R with
a bit offset of 0 for the current thread.</p>
<p>If the current call frame is not the top call frame, call frame information
(see <a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>) is used to determine the
location description that holds the register for the current call frame and
current program location of the current thread. The resulting location
description L is pushed.</p>
<p><em>Note that if call frame information is used, the resulting location
description may be register, memory, or undefined.</em></p>
<p><em>An implementation may evaluate the call frame information immediately, or
may defer evaluation until L is accessed by an operation. If evaluation is
deferred, R and the current context can be recorded in L. When accessed, the
recorded context is used to evaluate the call frame information, not the
current context of the access operation.</em></p>
</li>
</ol>
<p><em>These operations obtain a register location. To fetch the contents of a
register, it is necessary to use</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_regval_type</span></code><em>, use one of the</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_breg*</span></code> <em>register-based addressing operations, or use</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_deref*</span></code>
<em>on a register location description.</em></p>
</div>
<div class="section" id="implicit-location-description-operations">
<span id="amdgpu-dwarf-implicit-location-descriptions"></span><h7><a class="toc-backref" href="#id26">Implicit Location Description Operations</a><a class="headerlink" href="#implicit-location-description-operations" title="Permalink to this headline">¶</a></h7>
<p>Implicit location storage represents a piece or all of an object which has no
actual location in the program but whose contents are nonetheless known, either
as a constant or can be computed from other locations and values in the program.</p>
<p>An implicit location description specifies an implicit location storage. The bit
offset corresponds to a bit position within the implicit location storage. Bits
accessed using an implicit location description, access the corresponding
implicit storage value starting at the bit offset.</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_value</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_value</span></code> has two operands. The first is an unsigned LEB128
integer that represents a byte size S. The second is a block of bytes with a
length equal to S treated as a literal value V.</p>
<p>An implicit location storage LS is created with the literal value V and a
size of S.</p>
<p>It pushes location description L with one implicit location description SL
on the stack. SL specifies LS with a bit offset of 0.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_stack_value</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_stack_value</span></code> pops one stack entry that must be a value V.</p>
<p>An implicit location storage LS is created with the literal value V using
the size, encoding, and enianity specified by V’s base type.</p>
<p>It pushes a location description L with one implicit location description SL
on the stack. SL specifies LS with a bit offset of 0.</p>
<p><em>The</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_stack_value</span></code> <em>operation specifies that the object does not
exist in memory, but its value is nonetheless known. In this form, the
location description specifies the actual value of the object, rather than
specifying the memory or register storage that holds the value.</em></p>
<p>See <a class="reference internal" href="#amdgpu-dwarf-implicit-location-descriptions"><span class="std std-ref">Implicit Location Description Operations</span></a> for special rules
concerning implicit pointer values produced by dereferencing implicit
location descriptions created by the <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> and
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_implicit_aspace_pointer</span></code> operations.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Since location descriptions are allowed on the stack, the
<code class="docutils literal notranslate"><span class="pre">DW_OP_stack_value</span></code> operation no longer terminates the DWARF operation
expression execution as in DWARF Version 5.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code></p>
<p><em>An optimizing compiler may eliminate a pointer, while still retaining the
value that the pointer addressed.</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> <em>allows a
producer to describe this value.</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> <em>specifies an object is a pointer to the target
architecture default address space that cannot be represented as a real
pointer, even though the value it would point to can be described. In this
form, the location description specifies a debugging information entry that
represents the actual location description of the object to which the
pointer would point. Thus, a consumer of the debug information would be able
to access the dereferenced pointer, even when it cannot access the pointer
itself.</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> has two operands. The first operand is a 4-byte
unsigned value in the 32-bit DWARF format, or an 8-byte unsigned value in
the 64-bit DWARF format, that represents the byte offset DR of a debugging
information entry D relative to the beginning of the <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code> section
that contains the current compilation unit. The second operand is a signed
LEB128 integer that represents a byte displacement B.</p>
<p><em>Note that D may not be in the current compilation unit.</em></p>
<p><em>The first operand interpretation is exactly like that for</em>
<code class="docutils literal notranslate"><span class="pre">DW_FORM_ref_addr</span></code><em>.</em></p>
<p>The address space identifier AS is defined as the one corresponding to the
target architecture specific default address space.</p>
<p>The address size S is defined as the address bit size of the target
architecture specific address space corresponding to AS.</p>
<p>An implicit location storage LS is created with the debugging information
entry D, address space AS, and size of S.</p>
<p>It pushes a location description L that comprises one implicit location
description SL on the stack. SL specifies LS with a bit offset of 0.</p>
<p>It is an evaluation error if a <code class="docutils literal notranslate"><span class="pre">DW_OP_deref*</span></code> operation pops a location
description L’, and retrieves S bits, such that any retrieved bits come from
an implicit location storage that is the same as LS, unless both the
following conditions are met:</p>
<ol class="arabic">
<li><p>All retrieved bits come from an implicit location description that
refers to an implicit location storage that is the same as LS.</p>
<p><em>Note that all bits do not have to come from the same implicit location
description, as L’ may involve composite location descriptors.</em></p>
</li>
<li><p>The bits come from consecutive ascending offsets within their respective
implicit location storage.</p></li>
</ol>
<p><em>These rules are equivalent to retrieving the complete contents of LS.</em></p>
<p>If both the above conditions are met, then the value V pushed by the
<code class="docutils literal notranslate"><span class="pre">DW_OP_deref*</span></code> operation is an implicit pointer value IPV with a target
architecture specific address space of AS, a debugging information entry of
D, and a base type of T. If AS is the target architecture default address
space, then T is the generic type. Otherwise, T is a target architecture
specific integral type with a bit size equal to S.</p>
<p>If IPV is either implicitly converted to a location description (only done
if AS is the target architecture default address space) or used by
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code> (only done if the address space popped by
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code> is AS), then the resulting location
description RL is:</p>
<ul>
<li><p>If D has a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute, the DWARF expression E from the
<code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute is evaluated with the current context, except
that the result kind is a location description, the compilation unit is
the one that contains D, the object is unspecified, and the initial stack
is empty. RL is the expression result.</p>
<p><em>Note that E is evaluated with the context of the expression accessing
IPV, and not the context of the expression that contained the</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> <em>or</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code>
<em>operation that created L.</em></p>
</li>
<li><p>If D has a <code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code> attribute, then an implicit location
storage RLS is created from the <code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code> attribute’s value
with a size matching the size of the <code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code> attribute’s
value. RL comprises one implicit location description SRL. SRL specifies
RLS with a bit offset of 0.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If using <code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code> for variables and formal parameters is
deprecated and instead <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> is used with an implicit
location description, then this rule would not be required.</p>
</div>
</li>
<li><p>Otherwise, it is an evaluation error.</p></li>
</ul>
<p>The bit offset of RL is updated as if the <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset_uconst</span> <span class="pre">B</span></code>
operation was applied.</p>
<p>If a <code class="docutils literal notranslate"><span class="pre">DW_OP_stack_value</span></code> operation pops a value that is the same as IPV,
then it pushes a location description that is the same as L.</p>
<p>It is an evaluation error if LS or IPV is accessed in any other manner.</p>
<p><em>The restrictions on how an implicit pointer location description created
by</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> <em>and</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code>
<em>can be used are to simplify the DWARF consumer. Similarly, for an implicit
pointer value created by</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_deref*</span></code> <em>and</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_stack_value</span></code>.*</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code> has two operands that are the same as
for <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code>.</p>
<p>It pops one stack entry that must be an integral type value that represents
a target architecture specific address space identifier AS.</p>
<p>The location description L that is pushed on the stack is the same as for
<code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code>, except that the address space identifier used is
AS.</p>
<p>The DWARF expression is ill-formed if AS is not one of the values defined by
the target architecture specific <code class="docutils literal notranslate"><span class="pre">DW_ASPACE_*</span></code> values.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This definition of <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code> may change when
full support for address classes is added as required for languages such
as OpenCL/SyCL.</p>
</div>
</li>
</ol>
<p><em>Typically a</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> <em>or</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code> <em>operation is used in a DWARF expression
E</em><sub>1</sub> <em>of a</em> <code class="docutils literal notranslate"><span class="pre">DW_TAG_variable</span></code> <em>or</em> <code class="docutils literal notranslate"><span class="pre">DW_TAG_formal_parameter</span></code>
<em>debugging information entry D</em><sub>1</sub><em>‘s</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> <em>attribute.
The debugging information entry referenced by the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code>
<em>or</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code> <em>operations is typically itself a</em>
<code class="docutils literal notranslate"><span class="pre">DW_TAG_variable</span></code> <em>or</em> <code class="docutils literal notranslate"><span class="pre">DW_TAG_formal_parameter</span></code> <em>debugging information
entry D</em><sub>2</sub> <em>whose</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> <em>attribute gives a second DWARF
expression E</em><sub>2</sub><em>.</em></p>
<p><em>D</em><sub>1</sub> <em>and E</em><sub>1</sub> <em>are describing the location of a pointer type
object. D</em><sub>2</sub> <em>and E</em><sub>2</sub> <em>are describing the location of the
object pointed to by that pointer object.</em></p>
<p><em>However, D</em><sub>2</sub> <em>may be any debugging information entry that contains a</em>
<code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> <em>or</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code> <em>attribute (for example,</em>
<code class="docutils literal notranslate"><span class="pre">DW_TAG_dwarf_procedure</span></code><em>). By using E</em><sub>2</sub><em>, a consumer can
reconstruct the value of the object when asked to dereference the pointer
described by E</em><sub>1</sub> <em>which contains the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> or
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code> <em>operation.</em></p>
</div>
<div class="section" id="composite-location-description-operations">
<span id="amdgpu-dwarf-composite-location-description-operations"></span><h7><a class="toc-backref" href="#id27">Composite Location Description Operations</a><a class="headerlink" href="#composite-location-description-operations" title="Permalink to this headline">¶</a></h7>
<p>A composite location storage represents an object or value which may be
contained in part of another location storage or contained in parts of more
than one location storage.</p>
<p>Each part has a part location description L and a part bit size S. L can have
one or more single location descriptions SL. If there are more than one SL then
that indicates that part is located in more than one place. The bits of each
place of the part comprise S contiguous bits from the location storage LS
specified by SL starting at the bit offset specified by SL. All the bits must
be within the size of LS or the DWARF expression is ill-formed.</p>
<p>A composite location storage can have zero or more parts. The parts are
contiguous such that the zero-based location storage bit index will range over
each part with no gaps between them. Therefore, the size of a composite location
storage is the sum of the size of its parts. The DWARF expression is ill-formed
if the size of the contiguous location storage is larger than the size of the
memory location storage corresponding to the largest target architecture
specific address space.</p>
<p>A composite location description specifies a composite location storage. The bit
offset corresponds to a bit position within the composite location storage.</p>
<p>There are operations that create a composite location storage.</p>
<p>There are other operations that allow a composite location storage to be
incrementally created. Each part is created by a separate operation. There may
be one or more operations to create the final composite location storage. A
series of such operations describes the parts of the composite location storage
that are in the order that the associated part operations are executed.</p>
<p>To support incremental creation, a composite location storage can be in an
incomplete state. When an incremental operation operates on an incomplete
composite location storage, it adds a new part, otherwise it creates a new
composite location storage. The <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_piece_end</span></code> operation explicitly
makes an incomplete composite location storage complete.</p>
<p>A composite location description that specifies a composite location storage
that is incomplete is termed an incomplete composite location description. A
composite location description that specifies a composite location storage that
is complete is termed a complete composite location description.</p>
<p>If the top stack entry is a location description that has one incomplete
composite location description SL after the execution of an operation expression
has completed, SL is converted to a complete composite location description.</p>
<p><em>Note that this conversion does not happen after the completion of an operation
expression that is evaluated on the same stack by the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code>
<em>operations. Such executions are not a separate evaluation of an operation
expression, but rather the continued evaluation of the same operation expression
that contains the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> <em>operation.</em></p>
<p>If a stack entry is required to be a location description L, but L has an
incomplete composite location description, then the DWARF expression is
ill-formed. The exception is for the operations involved in incrementally
creating a composite location description as described below.</p>
<p><em>Note that a DWARF operation expression may arbitrarily compose composite
location descriptions from any other location description, including those that
have multiple single location descriptions, and those that have composite
location descriptions.</em></p>
<p><em>The incremental composite location description operations are defined to be
compatible with the definitions in DWARF Version 5.</em></p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_piece</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_piece</span></code> has a single unsigned LEB128 integer that represents a byte
size S.</p>
<p>The action is based on the context:</p>
<ul>
<li><p>If the stack is empty, then a location description L comprised of one
incomplete composite location description SL is pushed on the stack.</p>
<p>An incomplete composite location storage LS is created with a single part
P. P specifies a location description PL and has a bit size of S scaled by
8 (the byte size). PL is comprised of one undefined location description
PSL.</p>
<p>SL specifies LS with a bit offset of 0.</p>
</li>
<li><p>Otherwise, if the top stack entry is a location description L comprised of
one incomplete composite location description SL, then the incomplete
composite location storage LS that SL specifies is updated to append a new
part P. P specifies a location description PL and has a bit size of S
scaled by 8 (the byte size). PL is comprised of one undefined location
description PSL. L is left on the stack.</p></li>
<li><p>Otherwise, if the top stack entry is a location description or can be
converted to one, then it is popped and treated as a part location
description PL. Then:</p>
<ul>
<li><p>If the top stack entry (after popping PL) is a location description L
comprised of one incomplete composite location description SL, then the
incomplete composite location storage LS that SL specifies is updated to
append a new part P. P specifies the location description PL and has a
bit size of S scaled by 8 (the byte size). L is left on the stack.</p></li>
<li><p>Otherwise, a location description L comprised of one incomplete
composite location description SL is pushed on the stack.</p>
<p>An incomplete composite location storage LS is created with a single
part P. P specifies the location description PL and has a bit size of S
scaled by 8 (the byte size).</p>
<p>SL specifies LS with a bit offset of 0.</p>
</li>
</ul>
</li>
<li><p>Otherwise, the DWARF expression is ill-formed</p></li>
</ul>
<p><em>Many compilers store a single variable in sets of registers or store a
variable partially in memory and partially in registers.</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_piece</span></code>
<em>provides a way of describing where a part of a variable is located.</em></p>
<p><em>If a non-0 byte displacement is required, the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span></code>
<em>operation can be used to update the location description before using it as
the part location description of a</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_piece</span></code> <em>operation.</em></p>
<p><em>The evaluation rules for the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_piece</span></code> <em>operation allow it to be
compatible with the DWARF Version 5 definition.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Since these extensions allow location descriptions to be entries on the
stack, a simpler operation to create composite location descriptions could
be defined. For example, just one operation that specifies how many parts,
and pops pairs of stack entries for the part size and location
description. Not only would this be a simpler operation and avoid the
complexities of incomplete composite location descriptions, but it may
also have a smaller encoding in practice. However, the desire for
compatibility with DWARF Version 5 is likely a stronger consideration.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_bit_piece</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_bit_piece</span></code> has two operands. The first is an unsigned LEB128
integer that represents the part bit size S. The second is an unsigned
LEB128 integer that represents a bit displacement B.</p>
<p>The action is the same as for <code class="docutils literal notranslate"><span class="pre">DW_OP_piece</span></code>, except that any part created
has the bit size S, and the location description PL of any created part is
updated as if the <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">B;</span> <span class="pre">DW_OP_LLVM_bit_offset</span></code> operations were
applied.</p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_bit_piece</span></code> <em>is used instead of</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_piece</span></code> <em>when the piece to
be assembled is not byte-sized or is not at the start of the part location
description.</em></p>
<p><em>If a computed bit displacement is required, the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_bit_offset</span></code>
<em>operation can be used to update the location description before using it as
the part location description of a</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_bit_piece</span></code> <em>operation.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The bit offset operand is not needed as <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_bit_offset</span></code> can be
used on the part’s location description.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_piece_end</span></code> <em>New</em></p>
<p>If the top stack entry is not a location description L comprised of one
incomplete composite location description SL, then the DWARF expression is
ill-formed.</p>
<p>Otherwise, the incomplete composite location storage LS specified by SL is
updated to be a complete composite location description with the same parts.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_extend</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_extend</span></code> has two operands. The first is an unsigned LEB128
integer that represents the element bit size S. The second is an unsigned
LEB128 integer that represents a count C.</p>
<p>It pops one stack entry that must be a location description and is treated
as the part location description PL.</p>
<p>A location description L comprised of one complete composite location
description SL is pushed on the stack.</p>
<p>A complete composite location storage LS is created with C identical parts
P. Each P specifies PL and has a bit size of S.</p>
<p>SL specifies LS with a bit offset of 0.</p>
<p>The DWARF expression is ill-formed if the element bit size or count are 0.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_select_bit_piece</span></code> <em>New</em></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_select_bit_piece</span></code> has two operands. The first is an unsigned
LEB128 integer that represents the element bit size S. The second is an
unsigned LEB128 integer that represents a count C.</p>
<p>It pops three stack entries. The first must be an integral type value that
represents a bit mask value M. The second must be a location description
that represents the one-location description L1. The third must be a
location description that represents the zero-location description L0.</p>
<p>A complete composite location storage LS is created with C parts P<sub>N</sub>
ordered in ascending N from 0 to C-1 inclusive. Each P<sub>N</sub> specifies
location description PL<sub>N</sub> and has a bit size of S.</p>
<p>PL<sub>N</sub> is as if the <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_bit_offset</span> <span class="pre">N*S</span></code> operation was
applied to PLX<sub>N</sub>.</p>
<p>PLX<sub>N</sub> is the same as L0 if the N<sup>th</sup> least significant bit of
M is a zero, otherwise it is the same as L1.</p>
<p>A location description L comprised of one complete composite location
description SL is pushed on the stack. SL specifies LS with a bit offset of
0.</p>
<p>The DWARF expression is ill-formed if S or C are 0, or if the bit size of M
is less than C.</p>
</li>
</ol>
</div>
</div>
</div>
<div class="section" id="dwarf-location-list-expressions">
<span id="amdgpu-dwarf-location-list-expressions"></span><h5><a class="toc-backref" href="#id28">DWARF Location List Expressions</a><a class="headerlink" href="#dwarf-location-list-expressions" title="Permalink to this headline">¶</a></h5>
<p><em>To meet the needs of recent computer architectures and optimization techniques,
debugging information must be able to describe the location of an object whose
location changes over the object’s lifetime, and may reside at multiple
locations during parts of an object’s lifetime. Location list expressions are
used in place of operation expressions whenever the object whose location is
being described has these requirements.</em></p>
<p>A location list expression consists of a series of location list entries. Each
location list entry is one of the following kinds:</p>
<p><em>Bounded location description</em></p>
<blockquote>
<div><p>This kind of location list entry provides an operation expression that
evaluates to the location description of an object that is valid over a
lifetime bounded by a starting and ending address. The starting address is the
lowest address of the address range over which the location is valid. The
ending address is the address of the first location past the highest address
of the address range.</p>
<p>The location list entry matches when the current program location is within
the given range.</p>
<p>There are several kinds of bounded location description entries which differ
in the way that they specify the starting and ending addresses.</p>
</div></blockquote>
<p><em>Default location description</em></p>
<blockquote>
<div><p>This kind of location list entry provides an operation expression that
evaluates to the location description of an object that is valid when no
bounded location description entry applies.</p>
<p>The location list entry matches when the current program location is not
within the range of any bounded location description entry.</p>
</div></blockquote>
<p><em>Base address</em></p>
<blockquote>
<div><p>This kind of location list entry provides an address to be used as the base
address for beginning and ending address offsets given in certain kinds of
bounded location description entries. The applicable base address of a bounded
location description entry is the address specified by the closest preceding
base address entry in the same location list. If there is no preceding base
address entry, then the applicable base address defaults to the base address
of the compilation unit (see DWARF Version 5 section 3.1.1).</p>
<p>In the case of a compilation unit where all of the machine code is contained
in a single contiguous section, no base address entry is needed.</p>
</div></blockquote>
<p><em>End-of-list</em></p>
<blockquote>
<div><p>This kind of location list entry marks the end of the location list
expression.</p>
</div></blockquote>
<p>The address ranges defined by the bounded location description entries of a
location list expression may overlap. When they do, they describe a situation in
which an object exists simultaneously in more than one place.</p>
<p>If all of the address ranges in a given location list expression do not
collectively cover the entire range over which the object in question is
defined, and there is no following default location description entry, it is
assumed that the object is not available for the portion of the range that is
not covered.</p>
<p>The result of the evaluation of a DWARF location list expression is:</p>
<ul>
<li><p>If the current program location is not specified, then it is an evaluation
error.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If the location list only has a single default entry, should that be
considered a match if there is no program location? If there are non-default
entries then it seems it has to be an evaluation error when there is no
program location as that indicates the location depends on the program
location which is not known.</p>
</div>
</li>
<li><p>If there are no matching location list entries, then the result is a location
description that comprises one undefined location description.</p></li>
<li><p>Otherwise, the operation expression E of each matching location list entry is
evaluated with the current context, except that the result kind is a location
description, the object is unspecified, and the initial stack is empty. The
location list entry result is the location description returned by the
evaluation of E.</p>
<p>The result is a location description that is comprised of the union of the
single location descriptions of the location description result of each
matching location list entry.</p>
</li>
</ul>
<p>A location list expression can only be used as the value of a debugger
information entry attribute that is encoded using class <code class="docutils literal notranslate"><span class="pre">loclist</span></code> or
<code class="docutils literal notranslate"><span class="pre">loclistsptr</span></code> (see DWARF Version 5 section 7.5.5). The value of the attribute
provides an index into a separate object file section called <code class="docutils literal notranslate"><span class="pre">.debug_loclists</span></code>
or <code class="docutils literal notranslate"><span class="pre">.debug_loclists.dwo</span></code> (for split DWARF object files) that contains the
location list entries.</p>
<p>A <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> operation can be used to
specify a debugger information entry attribute that has a location list
expression. Several debugger information entry attributes allow DWARF
expressions that are evaluated with an initial stack that includes a location
description that may originate from the evaluation of a location list
expression.</p>
<p><em>This location list representation, the</em> <code class="docutils literal notranslate"><span class="pre">loclist</span></code> <em>and</em> <code class="docutils literal notranslate"><span class="pre">loclistsptr</span></code>
<em>class, and the related</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_loclists_base</span></code> <em>attribute are new in DWARF
Version 5. Together they eliminate most, or all of the code object relocations
previously needed for location list expressions.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The rest of this section is the same as DWARF Version 5 section 2.6.2.</p>
</div>
</div>
</div>
<div class="section" id="segmented-addresses">
<span id="amdgpu-dwarf-segment-addresses"></span><h4><a class="toc-backref" href="#id29">Segmented Addresses</a><a class="headerlink" href="#segmented-addresses" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 2.12.</p>
</div>
<p>DWARF address classes are used for source languages that have the concept of
memory spaces. They are used in the <code class="docutils literal notranslate"><span class="pre">DW_AT_address_class</span></code> attribute for
pointer type, reference type, subprogram, and subprogram type debugger
information entries.</p>
<p>Each DWARF address class is conceptually a separate source language memory space
with its own lifetime and aliasing rules. DWARF address classes are used to
specify the source language memory spaces that pointer type and reference type
values refer, and to specify the source language memory space in which variables
are allocated.</p>
<p>The set of currently defined source language DWARF address classes, together
with source language mappings, is given in
<a class="reference internal" href="#amdgpu-dwarf-address-class-table"><span class="std std-ref">Address class</span></a>.</p>
<p>Vendor defined source language address classes may be defined using codes in the
range <code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_lo_user</span></code> to <code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_hi_user</span></code>.</p>
<table class="docutils align-default" id="amdgpu-dwarf-address-class-table">
<caption><span class="caption-text">Address class</span><a class="headerlink" href="#amdgpu-dwarf-address-class-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 39%" />
<col style="width: 19%" />
<col style="width: 14%" />
<col style="width: 14%" />
<col style="width: 14%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Address Class Name</p></th>
<th class="head"><p>Meaning</p></th>
<th class="head"><p>C/C++</p></th>
<th class="head"><p>OpenCL</p></th>
<th class="head"><p>CUDA/HIP</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_none</span></code></p></td>
<td><p>generic</p></td>
<td><p><em>default</em></p></td>
<td><p>generic</p></td>
<td><p><em>default</em></p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_global</span></code></p></td>
<td><p>global</p></td>
<td></td>
<td><p>global</p></td>
<td></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_constant</span></code></p></td>
<td><p>constant</p></td>
<td></td>
<td><p>constant</p></td>
<td><p>constant</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_group</span></code></p></td>
<td><p>thread-group</p></td>
<td></td>
<td><p>local</p></td>
<td><p>shared</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_private</span></code></p></td>
<td><p>thread</p></td>
<td></td>
<td><p>private</p></td>
<td></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_lo_user</span></code></p></td>
<td></td>
<td></td>
<td></td>
<td></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_hi_user</span></code></p></td>
<td></td>
<td></td>
<td></td>
<td></td>
</tr>
</tbody>
</table>
<p>DWARF address spaces correspond to target architecture specific linear
addressable memory areas. They are used in DWARF expression location
descriptions to describe in which target architecture specific memory area data
resides.</p>
<p><em>Target architecture specific DWARF address spaces may correspond to hardware
supported facilities such as memory utilizing base address registers, scratchpad
memory, and memory with special interleaving. The size of addresses in these
address spaces may vary. Their access and allocation may be hardware managed
with each thread or group of threads having access to independent storage. For
these reasons they may have properties that do not allow them to be viewed as
part of the unified global virtual address space accessible by all threads.</em></p>
<p><em>It is target architecture specific whether multiple DWARF address spaces are
supported and how source language DWARF address classes map to target
architecture specific DWARF address spaces. A target architecture may map
multiple source language DWARF address classes to the same target architecture
specific DWARF address class. Optimization may determine that variable lifetime
and access pattern allows them to be allocated in faster scratchpad memory
represented by a different DWARF address space.</em></p>
<p>Although DWARF address space identifiers are target architecture specific,
<code class="docutils literal notranslate"><span class="pre">DW_ASPACE_none</span></code> is a common address space supported by all target
architectures.</p>
<p>DWARF address space identifiers are used by:</p>
<ul class="simple">
<li><p>The DWARF expression operations: <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_bregx</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_implicit_aspace_pointer</span></code>,
and <code class="docutils literal notranslate"><span class="pre">DW_OP_xderef*</span></code>.</p></li>
<li><p>The CFI instructions: <code class="docutils literal notranslate"><span class="pre">DW_CFA_LLVM_def_aspace_cfa</span></code> and
<code class="docutils literal notranslate"><span class="pre">DW_CFA_LLVM_def_aspace_cfa_sf</span></code>.</p></li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>With the definition of DWARF address classes and DWARF address spaces in these
extensions, DWARF Version 5 table 2.7 needs to be updated. It seems it is an
example of DWARF address spaces and not DWARF address classes.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>With the expanded support for DWARF address spaces in these extensions, it may
be worth examining if DWARF segments can be eliminated and DWARF address
spaces used instead.</p>
<p>That may involve extending DWARF address spaces to also be used to specify
code locations. In target architectures that use different memory areas for
code and data this would seem a natural use for DWARF address spaces. This
would allow DWARF expression location descriptions to be used to describe the
location of subprograms and entry points that are used in expressions
involving subprogram pointer type values.</p>
<p>Currently, DWARF expressions assume data and code resides in the same default
DWARF address space, and only the address ranges in DWARF location list
entries and in the <code class="docutils literal notranslate"><span class="pre">.debug_aranges</span></code> section for accelerated access for
addresses allow DWARF segments to be used to distinguish.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Currently, DWARF defines address class values as being target architecture
specific. It is unclear how language specific memory spaces are intended to be
represented in DWARF using these.</p>
<p>For example, OpenCL defines memory spaces (called address spaces in OpenCL)
for <code class="docutils literal notranslate"><span class="pre">global</span></code>, <code class="docutils literal notranslate"><span class="pre">local</span></code>, <code class="docutils literal notranslate"><span class="pre">constant</span></code>, and <code class="docutils literal notranslate"><span class="pre">private</span></code>. These are part of
the type system and are modifiers to pointer types. In addition, OpenCL
defines <code class="docutils literal notranslate"><span class="pre">generic</span></code> pointers that can reference either the <code class="docutils literal notranslate"><span class="pre">global</span></code>,
<code class="docutils literal notranslate"><span class="pre">local</span></code>, or <code class="docutils literal notranslate"><span class="pre">private</span></code> memory spaces. To support the OpenCL language the
debugger would want to support casting pointers between the <code class="docutils literal notranslate"><span class="pre">generic</span></code> and
other memory spaces, querying what memory space a <code class="docutils literal notranslate"><span class="pre">generic</span></code> pointer value is
currently referencing, and possibly using pointer casting to form an address
for a specific memory space out of an integral value.</p>
<p>The method to use to dereference a pointer type or reference type value is
defined in DWARF expressions using <code class="docutils literal notranslate"><span class="pre">DW_OP_xderef*</span></code> which uses a target
architecture specific address space.</p>
<p>DWARF defines the <code class="docutils literal notranslate"><span class="pre">DW_AT_address_class</span></code> attribute on pointer type and
reference type debugger information entries. It specifies the method to use to
dereference them. Why is the value of this not the same as the address space
value used in <code class="docutils literal notranslate"><span class="pre">DW_OP_xderef*</span></code>? In both cases it is target architecture
specific and the architecture presumably will use the same set of methods to
dereference pointers in both cases.</p>
<p>Since <code class="docutils literal notranslate"><span class="pre">DW_AT_address_class</span></code> uses a target architecture specific value, it
cannot in general capture the source language memory space type modifier
concept. On some architectures all source language memory space modifiers may
actually use the same method for dereferencing pointers.</p>
<p>One possibility is for DWARF to add an <code class="docutils literal notranslate"><span class="pre">DW_TAG_LLVM_address_class_type</span></code>
debugger information entry type modifier that can be applied to a pointer type
and reference type. The <code class="docutils literal notranslate"><span class="pre">DW_AT_address_class</span></code> attribute could be re-defined
to not be target architecture specific and instead define generalized language
values (as presented above for DWARF address classes in the table
<a class="reference internal" href="#amdgpu-dwarf-address-class-table"><span class="std std-ref">Address class</span></a>) that will support OpenCL and other
languages using memory spaces. The <code class="docutils literal notranslate"><span class="pre">DW_AT_address_class</span></code> attribute could be
defined to not be applied to pointer types or reference types, but instead
only to the new <code class="docutils literal notranslate"><span class="pre">DW_TAG_LLVM_address_class_type</span></code> type modifier debugger
information entry.</p>
<p>If a pointer type or reference type is not modified by
<code class="docutils literal notranslate"><span class="pre">DW_TAG_LLVM_address_class_type</span></code> or if <code class="docutils literal notranslate"><span class="pre">DW_TAG_LLVM_address_class_type</span></code>
has no <code class="docutils literal notranslate"><span class="pre">DW_AT_address_class</span></code> attribute, then the pointer type or reference
type would be defined to use the <code class="docutils literal notranslate"><span class="pre">DW_ADDR_none</span></code> address class as currently.
Since modifiers can be chained, it would need to be defined if multiple
<code class="docutils literal notranslate"><span class="pre">DW_TAG_LLVM_address_class_type</span></code> modifiers were legal, and if so if the
outermost one is the one that takes precedence.</p>
<p>A target architecture implementation that supports multiple address spaces
would need to map <code class="docutils literal notranslate"><span class="pre">DW_ADDR_none</span></code> appropriately to support CUDA-like
languages that have no address classes in the type system but do support
variable allocation in address classes. Such variable allocation would result
in the variable’s location description needing an address space.</p>
<p>The approach presented in <a class="reference internal" href="#amdgpu-dwarf-address-class-table"><span class="std std-ref">Address class</span></a> is to define
the default <code class="docutils literal notranslate"><span class="pre">DW_ADDR_none</span></code> to be the generic address class and not the
global address class. This matches how CLANG and LLVM have added support for
CUDA-like languages on top of existing C++ language support. This allows all
addresses to be generic by default which matches CUDA-like languages.</p>
<p>An alternative approach is to define <code class="docutils literal notranslate"><span class="pre">DW_ADDR_none</span></code> as being the global
address class and then change <code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_global</span></code> to
<code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_generic</span></code>. This would match the reality that languages that do
not support multiple memory spaces only have one default global memory space.
Generally, in these languages if they expose that the target architecture
supports multiple address spaces, the default one is still the global memory
space. Then a language that does support multiple memory spaces has to
explicitly indicate which pointers have the added ability to reference more
than the global memory space. However, compilers generating DWARF for
CUDA-like languages would then have to define every CUDA-like language pointer
type or reference type using <code class="docutils literal notranslate"><span class="pre">DW_TAG_LLVM_address_class_type</span></code> with a
<code class="docutils literal notranslate"><span class="pre">DW_AT_address_class</span></code> attribute of <code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_generic</span></code> to match the
language semantics.</p>
<p>A new <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_address_space</span></code> attribute could be defined that can be
applied to pointer type, reference type, subprogram, and subprogram type to
describe how objects having the given type are dereferenced or called (the
role that <code class="docutils literal notranslate"><span class="pre">DW_AT_address_class</span></code> currently provides). The values of
<code class="docutils literal notranslate"><span class="pre">DW_AT_address_space</span></code> would be target architecture specific and the same as
used in <code class="docutils literal notranslate"><span class="pre">DW_OP_xderef*</span></code>.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Some additional changes will be made to support languages such as OpenCL/SyCL
that allow address class pointer casting and queries.</p>
<p>This requires the compiler to provide the mapping from address space to
address class which may be runtime and not target architecture dependent. Some
implementations may have a one-to-one mapping from source language address
class to target architecture address space, and some may have a many-to-one
mapping which requires knowledge of the address class when determining if
pointer address class casts are allowed.</p>
<p>The changes will likely add an attribute that has an expression provided by
the compiler to map from address class to address space. The
<code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_implicit_pointer</span></code>
operations may be changed as the current IPV definition may not provide enough
information when used to cast between address classes. Other attributes and
operations may be needed. The legal casts between address classes may need to
be defined on a per language address class basis.</p>
</div>
</div>
<div class="section" id="debugging-information-entry-attributes">
<span id="amdgpu-dwarf-debugging-information-entry-attributes"></span><h4><a class="toc-backref" href="#id30">Debugging Information Entry Attributes</a><a class="headerlink" href="#debugging-information-entry-attributes" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This section provides changes to existing debugger information entry
attributes and defines attributes added by these extensions. These would be
incorporated into the appropriate DWARF Version 5 chapter 2 sections.</p>
</div>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code></p>
<p>Any debugging information entry describing a data object (which includes
variables and parameters) or common blocks may have a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code>
attribute, whose value is a DWARF expression E.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a location description, an unspecified object, the
compilation unit that contains E, an empty initial stack, and other context
elements corresponding to the source language thread of execution upon which
the user is focused, if any. The result of the evaluation is the location
description of the base of the data object.</p>
<p>See <a class="reference internal" href="#amdgpu-dwarf-control-flow-operations"><span class="std std-ref">Control Flow Operations</span></a> for special evaluation rules
used by the <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> operations.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Delete the description of how the <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> operations evaluate a
<code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute as that is now described in the operations.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>See the discussion about the <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute in the
<code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> operation. Having each attribute only have a single
purpose and single execution semantics seems desirable. It makes it easier
for the consumer that no longer have to track the context. It makes it
easier for the producer as it can rely on a single semantics for each
attribute.</p>
<p>For that reason, limiting the <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> attribute to only
supporting evaluating the location description of an object, and using a
different attribute and encoding class for the evaluation of DWARF
expression <em>procedures</em> on the same operation expression stack seems
desirable.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Could deprecate using the <code class="docutils literal notranslate"><span class="pre">DW_AT_const_value</span></code> attribute for
<code class="docutils literal notranslate"><span class="pre">DW_TAG_variable</span></code> or <code class="docutils literal notranslate"><span class="pre">DW_TAG_formal_parameter</span></code> debugger information
entries that have been optimized to a constant. Instead,
<code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code> could be used with a DWARF expression that produces an
implicit location description now that any location description can be
used within a DWARF expression. This allows the <code class="docutils literal notranslate"><span class="pre">DW_OP_call*</span></code> operations
to be used to push the location description of any variable regardless of
how it is optimized.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_frame_base</span></code></p>
<p>A <code class="docutils literal notranslate"><span class="pre">DW_TAG_subprogram</span></code> or <code class="docutils literal notranslate"><span class="pre">DW_TAG_entry_point</span></code> debugger information entry
may have a <code class="docutils literal notranslate"><span class="pre">DW_AT_frame_base</span></code> attribute, whose value is a DWARF expression
E.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a location description, an unspecified object, the
compilation unit that contains E, an empty initial stack, and other context
elements corresponding to the source language thread of execution upon which
the user is focused, if any.</p>
<p>The DWARF is ill-formed if E contains an <code class="docutils literal notranslate"><span class="pre">DW_OP_fbreg</span></code> operation, or the
resulting location description L is not comprised of one single location
description SL.</p>
<p>If SL a register location description for register R, then L is replaced
with the result of evaluating a <code class="docutils literal notranslate"><span class="pre">DW_OP_bregx</span> <span class="pre">R,</span> <span class="pre">0</span></code> operation. This
computes the frame base memory location description in the target
architecture default address space.</p>
<p><em>This allows the more compact</em> <code class="docutils literal notranslate"><span class="pre">DW_OPreg*</span></code> <em>to be used instead of</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_breg*</span> <span class="pre">0</span></code><em>.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This rule could be removed and require the producer to create the required
location description directly using <code class="docutils literal notranslate"><span class="pre">DW_OP_call_frame_cfa</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_OP_breg*</span></code>, or <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_aspace_bregx</span></code>. This would also then
allow a target to implement the call frames within a large register.</p>
</div>
<p>Otherwise, the DWARF is ill-formed if SL is not a memory location
description in any of the target architecture specific address spaces.</p>
<p>The resulting L is the <em>frame base</em> for the subprogram or entry point.</p>
<p><em>Typically, E will use the</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_call_frame_cfa</span></code> <em>operation or be a
stack pointer register plus or minus some offset.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_data_member_location</span></code></p>
<p>For a <code class="docutils literal notranslate"><span class="pre">DW_AT_data_member_location</span></code> attribute there are two cases:</p>
<ol class="arabic">
<li><p>If the attribute is an integer constant B, it provides the offset in
bytes from the beginning of the containing entity.</p>
<p>The result of the attribute is obtained by evaluating a
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span> <span class="pre">B</span></code> operation with an initial stack comprising the
location description of the beginning of the containing entity. The
result of the evaluation is the location description of the base of the
member entry.</p>
<p><em>If the beginning of the containing entity is not byte aligned, then the
beginning of the member entry has the same bit displacement within a
byte.</em></p>
</li>
<li><p>Otherwise, the attribute must be a DWARF expression E which is evaluated
with a context that has a result kind of a location description, an
unspecified object, the compilation unit that contains E, an initial
stack comprising the location description of the beginning of the
containing entity, and other context elements corresponding to the
source language thread of execution upon which the user is focused, if
any. The result of the evaluation is the location description of the
base of the member entry.</p></li>
</ol>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The beginning of the containing entity can now be any location
description, including those with more than one single location
description, and those with single location descriptions that are of any
kind and have any bit offset.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_use_location</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_TAG_ptr_to_member_type</span></code> debugging information entry has a
<code class="docutils literal notranslate"><span class="pre">DW_AT_use_location</span></code> attribute whose value is a DWARF expression E. It is
used to compute the location description of the member of the class to which
the pointer to member entry points.</p>
<p><em>The method used to find the location description of a given member of a
class, structure, or union is common to any instance of that class,
structure, or union and to any instance of the pointer to member type. The
method is thus associated with the pointer to member type, rather than with
each object that has a pointer to member type.</em></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_AT_use_location</span></code> DWARF expression is used in conjunction with the
location description for a particular object of the given pointer to member
type and for a particular structure or class instance.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a location description, an unspecified object, the
compilation unit that contains E, an initial stack comprising two entries,
and other context elements corresponding to the source language thread of
execution upon which the user is focused, if any. The first stack entry is
the value of the pointer to member object itself. The second stack entry is
the location description of the base of the entire class, structure, or
union instance containing the member whose location is being calculated. The
result of the evaluation is the location description of the member of the
class to which the pointer to member entry points.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_data_location</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_AT_data_location</span></code> attribute may be used with any type that
provides one or more levels of hidden indirection and/or run-time parameters
in its representation. Its value is a DWARF operation expression E which
computes the location description of the data for an object. When this
attribute is omitted, the location description of the data is the same as
the location description of the object.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a location description, an object that is the location
description of the data descriptor, the compilation unit that contains E, an
empty initial stack, and other context elements corresponding to the source
language thread of execution upon which the user is focused, if any. The
result of the evaluation is the location description of the base of the
member entry.</p>
<p><em>E will typically involve an operation expression that begins with a</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> <em>operation which loads the location
description of the object which can then serve as a description in
subsequent calculation.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Since <code class="docutils literal notranslate"><span class="pre">DW_AT_data_member_location</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_AT_use_location</span></code>, and
<code class="docutils literal notranslate"><span class="pre">DW_AT_vtable_elem_location</span></code> allow both operation expressions and
location list expressions, why does <code class="docutils literal notranslate"><span class="pre">DW_AT_data_location</span></code> not allow
both? In all cases they apply to data objects so less likely that
optimization would cause different operation expressions for different
program location ranges. But if supporting for some then should be for
all.</p>
<p>It seems odd this attribute is not the same as
<code class="docutils literal notranslate"><span class="pre">DW_AT_data_member_location</span></code> in having an initial stack with the
location description of the object since the expression has to need it.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_vtable_elem_location</span></code></p>
<p>An entry for a virtual function also has a <code class="docutils literal notranslate"><span class="pre">DW_AT_vtable_elem_location</span></code>
attribute whose value is a DWARF expression E.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a location description, an unspecified object, the
compilation unit that contains E, an initial stack comprising the location
description of the object of the enclosing type, and other context elements
corresponding to the source language thread of execution upon which the user
is focused, if any. The result of the evaluation is the location description
of the slot for the function within the virtual function table for the
enclosing class.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_static_link</span></code></p>
<p>If a <code class="docutils literal notranslate"><span class="pre">DW_TAG_subprogram</span></code> or <code class="docutils literal notranslate"><span class="pre">DW_TAG_entry_point</span></code> debugger information
entry is lexically nested, it may have a <code class="docutils literal notranslate"><span class="pre">DW_AT_static_link</span></code> attribute,
whose value is a DWARF expression E.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a location description, an unspecified object, the
compilation unit that contains E, an empty initial stack, and other context
elements corresponding to the source language thread of execution upon which
the user is focused, if any. The result of the evaluation is the location
description L of the <em>canonical frame address</em> (see
<a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a>) of the relevant call frame of
the subprogram instance that immediately lexically encloses the current call
frame’s subprogram or entry point.</p>
<p>The DWARF is ill-formed if L is is not comprised of one memory location
description for one of the target architecture specific address spaces.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_return_addr</span></code></p>
<p>A <code class="docutils literal notranslate"><span class="pre">DW_TAG_subprogram</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_TAG_inlined_subroutine</span></code>, or
<code class="docutils literal notranslate"><span class="pre">DW_TAG_entry_point</span></code> debugger information entry may have a
<code class="docutils literal notranslate"><span class="pre">DW_AT_return_addr</span></code> attribute, whose value is a DWARF expression E.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a location description, an unspecified object, the
compilation unit that contains E, an empty initial stack, and other context
elements corresponding to the source language thread of execution upon which
the user is focused, if any. The result of the evaluation is the location
description L of the place where the return address for the current call
frame’s subprogram or entry point is stored.</p>
<p>The DWARF is ill-formed if L is not comprised of one memory location
description for one of the target architecture specific address spaces.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>It is unclear why <code class="docutils literal notranslate"><span class="pre">DW_TAG_inlined_subroutine</span></code> has a
<code class="docutils literal notranslate"><span class="pre">DW_AT_return_addr</span></code> attribute but not a <code class="docutils literal notranslate"><span class="pre">DW_AT_frame_base</span></code> or
<code class="docutils literal notranslate"><span class="pre">DW_AT_static_link</span></code> attribute. Seems it would either have all of them or
none. Since inlined subprograms do not have a call frame it seems they
would have none of these attributes.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_call_value</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_location</span></code>, and
<code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_value</span></code></p>
<p>A <code class="docutils literal notranslate"><span class="pre">DW_TAG_call_site_parameter</span></code> debugger information entry may have a
<code class="docutils literal notranslate"><span class="pre">DW_AT_call_value</span></code> attribute, whose value is a DWARF operation expression
E<sub>1</sub>.</p>
<p>The result of the <code class="docutils literal notranslate"><span class="pre">DW_AT_call_value</span></code> attribute is obtained by evaluating
E<sub>1</sub> with a context that has a result kind of a value, an unspecified
object, the compilation unit that contains E, an empty initial stack, and
other context elements corresponding to the source language thread of
execution upon which the user is focused, if any. The resulting value V<sub>1</sub> is the value of the parameter at the time of the call made by the
call site.</p>
<p>For parameters passed by reference, where the code passes a pointer to a
location which contains the parameter, or for reference type parameters, the
<code class="docutils literal notranslate"><span class="pre">DW_TAG_call_site_parameter</span></code> debugger information entry may also have a
<code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_location</span></code> attribute whose value is a DWARF operation
expression E<sub>2</sub>, and a <code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_value</span></code> attribute whose
value is a DWARF operation expression E<sub>3</sub>.</p>
<p>The value of the <code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_location</span></code> attribute is obtained by
evaluating E<sub>2</sub> with a context that has a result kind of a location
description, an unspecified object, the compilation unit that contains E, an
empty initial stack, and other context elements corresponding to the source
language thread of execution upon which the user is focused, if any. The
resulting location description L<sub>2</sub> is the location where the
referenced parameter lives during the call made by the call site. If E<sub>2</sub> would just be a <code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code>, then the
<code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_location</span></code> attribute may be omitted.</p>
<p>The value of the <code class="docutils literal notranslate"><span class="pre">DW_AT_call_data_value</span></code> attribute is obtained by
evaluating E<sub>3</sub> with a context that has a result kind of a value, an
unspecified object, the compilation unit that contains E, an empty initial
stack, and other context elements corresponding to the source language
thread of execution upon which the user is focused, if any. The resulting
value V<sub>3</sub> is the value in L<sub>2</sub> at the time of the call made
by the call site.</p>
<p>The result of these attributes is undefined if the current call frame is
not for the subprogram containing the <code class="docutils literal notranslate"><span class="pre">DW_TAG_call_site_parameter</span></code>
debugger information entry or the current program location is not for the
call site containing the <code class="docutils literal notranslate"><span class="pre">DW_TAG_call_site_parameter</span></code> debugger information
entry in the current call frame.</p>
<p><em>The consumer may have to virtually unwind to the call site (see</em>
<a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a><em>) in order to evaluate these
attributes. This will ensure the source language thread of execution upon
which the user is focused corresponds to the call site needed to evaluate
the expression.</em></p>
<p>If it is not possible to avoid the expressions of these attributes from
accessing registers or memory locations that might be clobbered by the
subprogram being called by the call site, then the associated attribute
should not be provided.</p>
<p><em>The reason for the restriction is that the parameter may need to be
accessed during the execution of the callee. The consumer may virtually
unwind from the called subprogram back to the caller and then evaluate the
attribute expressions. The call frame information (see</em>
<a class="reference internal" href="#amdgpu-dwarf-call-frame-information"><span class="std std-ref">Call Frame Information</span></a><em>) will not be able to restore
registers that have been clobbered, and clobbered memory will no longer have
the value at the time of the call.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lanes</span></code> <em>New</em></p>
<p>For languages that are implemented using a SIMD or SIMT execution model, a
<code class="docutils literal notranslate"><span class="pre">DW_TAG_subprogram</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_TAG_inlined_subroutine</span></code>, or
<code class="docutils literal notranslate"><span class="pre">DW_TAG_entry_point</span></code> debugger information entry may have a
<code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lanes</span></code> attribute whose value is an integer constant that is
the number of lanes per thread. This is the static number of lanes per
thread. It is not the dynamic number of lanes with which the thread was
initiated, for example, due to smaller or partial work-groups.</p>
<p>If not present, the default value of 1 is used.</p>
<p>The DWARF is ill-formed if the value is 0.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code> <em>New</em></p>
<p>For languages that are implemented using a SIMD or SIMT execution model, a
<code class="docutils literal notranslate"><span class="pre">DW_TAG_subprogram</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_TAG_inlined_subroutine</span></code>, or
<code class="docutils literal notranslate"><span class="pre">DW_TAG_entry_point</span></code> debugging information entry may have a
<code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code> attribute whose value is a DWARF expression E.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a location description, an unspecified object, the
compilation unit that contains E, an empty initial stack, and other context
elements corresponding to the source language thread of execution upon which
the user is focused, if any.</p>
<p>The resulting location description L is for a thread lane count sized vector
of generic type elements. The thread lane count is the value of the
<code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lanes</span></code> attribute. Each element holds the conceptual program
location of the corresponding lane, where the least significant element
corresponds to the first target architecture specific lane identifier and so
forth. If the lane was not active when the current subprogram was called,
its element is an undefined location description.</p>
<p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code> <em>allows the compiler to indicate conceptually where
each lane of a SIMT thread is positioned even when it is in divergent
control flow that is not active.</em></p>
<p><em>Typically, the result is a location description with one composite location
description with each part being a location description with either one
undefined location description or one memory location description.</em></p>
<p>If not present, the thread is not being used in a SIMT manner, and the
thread’s current program location is used.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_active_lane</span></code> <em>New</em></p>
<p>For languages that are implemented using a SIMD or SIMT execution model, a
<code class="docutils literal notranslate"><span class="pre">DW_TAG_subprogram</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_TAG_inlined_subroutine</span></code>, or
<code class="docutils literal notranslate"><span class="pre">DW_TAG_entry_point</span></code> debugger information entry may have a
<code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_active_lane</span></code> attribute whose value is a DWARF expression E.</p>
<p>The result of the attribute is obtained by evaluating E with a context that
has a result kind of a value, an unspecified object, the compilation unit
that contains E, an empty initial stack, and other context elements
corresponding to the source language thread of execution upon which the user
is focused, if any.</p>
<p>The DWARF is ill-formed if the resulting value V is not an integral value.</p>
<p>The resulting V is a bit mask of active lanes for the current program
location. The N<sup>th</sup> least significant bit of the mask corresponds to
the N<sup>th</sup> lane. If the bit is 1 the lane is active, otherwise it is
inactive.</p>
<p><em>Some targets may update the target architecture execution mask for regions
of code that must execute with different sets of lanes than the current
active lanes. For example, some code must execute with all lanes made
temporarily active.</em> <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_active_lane</span></code> <em>allows the compiler to
provide the means to determine the source language active lanes.</em></p>
<p>If not present and <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lanes</span></code> is greater than 1, then the target
architecture execution mask is used.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_vector_size</span></code> <em>New</em></p>
<p>A <code class="docutils literal notranslate"><span class="pre">DW_TAG_base_type</span></code> debugger information entry for a base type T may have
a <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_vector_size</span></code> attribute whose value is an integer constant
that is the vector type size N.</p>
<p>The representation of a vector base type is as N contiguous elements, each
one having the representation of a base type T’ that is the same as T
without the <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_vector_size</span></code> attribute.</p>
<p>If a <code class="docutils literal notranslate"><span class="pre">DW_TAG_base_type</span></code> debugger information entry does not have a
<code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_vector_size</span></code> attribute, then the base type is not a vector
type.</p>
<p>The DWARF is ill-formed if N is not greater than 0.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>LLVM has mention of a non-upstreamed debugger information entry that is
intended to support vector types. However, that was not for a base type so
would not be suitable as the type of a stack value entry. But perhaps that
could be replaced by using this attribute.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_augmentation</span></code> <em>New</em></p>
<p>A <code class="docutils literal notranslate"><span class="pre">DW_TAG_compile_unit</span></code> debugger information entry for a compilation unit
may have a <code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_augmentation</span></code> attribute, whose value is an
augmentation string.</p>
<p><em>The augmentation string allows producers to indicate that there is
additional vendor or target specific information in the debugging
information entries. For example, this might be information about the
version of vendor specific extensions that are being used.</em></p>
<p>If not present, or if the string is empty, then the compilation unit has no
augmentation string.</p>
<p>The format for the augmentation string is:</p>
<blockquote>
<div><div class="line-block">
<div class="line"><code class="docutils literal notranslate"><span class="pre">[</span></code><em>vendor</em><code class="docutils literal notranslate"><span class="pre">:v</span></code><em>X</em><code class="docutils literal notranslate"><span class="pre">.</span></code><em>Y</em>[<code class="docutils literal notranslate"><span class="pre">:</span></code><em>options</em>]<code class="docutils literal notranslate"><span class="pre">]</span></code>*</div>
</div>
</div></blockquote>
<p>Where <em>vendor</em> is the producer, <code class="docutils literal notranslate"><span class="pre">vX.Y</span></code> specifies the major X and minor Y
version number of the extensions used, and <em>options</em> is an optional string
providing additional information about the extensions. The version number
must conform to semantic versioning [<a class="reference internal" href="#amdgpu-dwarf-semver"><span class="std std-ref">SEMVER</span></a>].
The <em>options</em> string must not contain the “<code class="docutils literal notranslate"><span class="pre">]</span></code>” character.</p>
<p>For example:</p>
<blockquote>
<div><div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="p">[</span><span class="n">abc</span><span class="p">:</span><span class="n">v0</span><span class="o">.</span><span class="mi">0</span><span class="p">][</span><span class="n">def</span><span class="p">:</span><span class="n">v1</span><span class="o">.</span><span class="mi">2</span><span class="p">:</span><span class="n">feature</span><span class="o">-</span><span class="n">a</span><span class="o">=</span><span class="n">on</span><span class="p">,</span><span class="n">feature</span><span class="o">-</span><span class="n">b</span><span class="o">=</span><span class="mi">3</span><span class="p">]</span>
</pre></div>
</div>
</div></blockquote>
</li>
</ol>
</div>
</div>
<div class="section" id="program-scope-entities">
<h3><a class="toc-backref" href="#id31">Program Scope Entities</a><a class="headerlink" href="#program-scope-entities" title="Permalink to this headline">¶</a></h3>
<div class="section" id="unit-entities">
<span id="amdgpu-dwarf-language-names"></span><h4><a class="toc-backref" href="#id32">Unit Entities</a><a class="headerlink" href="#unit-entities" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 3.1.1 and Table 3.1.</p>
</div>
<p>Additional language codes defined for use with the <code class="docutils literal notranslate"><span class="pre">DW_AT_language</span></code> attribute
are defined in <a class="reference internal" href="#amdgpu-dwarf-language-names-table"><span class="std std-ref">Language Names</span></a>.</p>
<table class="docutils align-default" id="amdgpu-dwarf-language-names-table">
<caption><span class="caption-text">Language Names</span><a class="headerlink" href="#amdgpu-dwarf-language-names-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 41%" />
<col style="width: 59%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Language Name</p></th>
<th class="head"><p>Meaning</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_LANG_LLVM_HIP</span></code></p></td>
<td><p>HIP Language.</p></td>
</tr>
</tbody>
</table>
<p>The HIP language [<a class="reference internal" href="#amdgpu-dwarf-hip"><span class="std std-ref">HIP</span></a>] can be supported by extending
the C++ language.</p>
</div>
</div>
<div class="section" id="other-debugger-information">
<h3><a class="toc-backref" href="#id33">Other Debugger Information</a><a class="headerlink" href="#other-debugger-information" title="Permalink to this headline">¶</a></h3>
<div class="section" id="accelerated-access">
<h4><a class="toc-backref" href="#id34">Accelerated Access</a><a class="headerlink" href="#accelerated-access" title="Permalink to this headline">¶</a></h4>
<div class="section" id="lookup-by-name">
<span id="amdgpu-dwarf-lookup-by-name"></span><h5><a class="toc-backref" href="#id35">Lookup By Name</a><a class="headerlink" href="#lookup-by-name" title="Permalink to this headline">¶</a></h5>
<div class="section" id="contents-of-the-name-index">
<h6><a class="toc-backref" href="#id36">Contents of the Name Index</a><a class="headerlink" href="#contents-of-the-name-index" title="Permalink to this headline">¶</a></h6>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The following provides changes to DWARF Version 5 section 6.1.1.1.</p>
<p>The rule for debugger information entries included in the name index in the
optional <code class="docutils literal notranslate"><span class="pre">.debug_names</span></code> section is extended to also include named
<code class="docutils literal notranslate"><span class="pre">DW_TAG_variable</span></code> debugging information entries with a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code>
attribute that includes a <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code> operation.</p>
</div>
<p>The name index must contain an entry for each debugging information entry that
defines a named subprogram, label, variable, type, or namespace, subject to the
following rules:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_TAG_variable</span></code> debugging information entries with a <code class="docutils literal notranslate"><span class="pre">DW_AT_location</span></code>
attribute that includes a <code class="docutils literal notranslate"><span class="pre">DW_OP_addr</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_form_aspace_address</span></code>,
or <code class="docutils literal notranslate"><span class="pre">DW_OP_form_tls_address</span></code> operation are included; otherwise, they are
excluded.</p></li>
</ul>
</div>
<div class="section" id="data-representation-of-the-name-index">
<h6><a class="toc-backref" href="#id37">Data Representation of the Name Index</a><a class="headerlink" href="#data-representation-of-the-name-index" title="Permalink to this headline">¶</a></h6>
<div class="section" id="section-header">
<h7><a class="toc-backref" href="#id38">Section Header</a><a class="headerlink" href="#section-header" title="Permalink to this headline">¶</a></h7>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The following provides an addition to DWARF Version 5 section 6.1.1.4.1 item
14 <code class="docutils literal notranslate"><span class="pre">augmentation_string</span></code>.</p>
</div>
<p>A null-terminated UTF-8 vendor specific augmentation string, which provides
additional information about the contents of this index. If provided, the
recommended format for augmentation string is:</p>
<blockquote>
<div><div class="line-block">
<div class="line"><code class="docutils literal notranslate"><span class="pre">[</span></code><em>vendor</em><code class="docutils literal notranslate"><span class="pre">:v</span></code><em>X</em><code class="docutils literal notranslate"><span class="pre">.</span></code><em>Y</em>[<code class="docutils literal notranslate"><span class="pre">:</span></code><em>options</em>]<code class="docutils literal notranslate"><span class="pre">]</span></code>*</div>
</div>
</div></blockquote>
<p>Where <em>vendor</em> is the producer, <code class="docutils literal notranslate"><span class="pre">vX.Y</span></code> specifies the major X and minor Y
version number of the extensions used in the DWARF of the compilation unit, and
<em>options</em> is an optional string providing additional information about the
extensions. The version number must conform to semantic versioning [<a class="reference internal" href="#amdgpu-dwarf-semver"><span class="std std-ref">SEMVER</span></a>]. The <em>options</em> string must not contain the “<code class="docutils literal notranslate"><span class="pre">]</span></code>”
character.</p>
<p>For example:</p>
<blockquote>
<div><div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="p">[</span><span class="n">abc</span><span class="p">:</span><span class="n">v0</span><span class="o">.</span><span class="mi">0</span><span class="p">][</span><span class="n">def</span><span class="p">:</span><span class="n">v1</span><span class="o">.</span><span class="mi">2</span><span class="p">:</span><span class="n">feature</span><span class="o">-</span><span class="n">a</span><span class="o">=</span><span class="n">on</span><span class="p">,</span><span class="n">feature</span><span class="o">-</span><span class="n">b</span><span class="o">=</span><span class="mi">3</span><span class="p">]</span>
</pre></div>
</div>
</div></blockquote>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This is different to the definition in DWARF Version 5 but is consistent with
the other augmentation strings and allows multiple vendor extensions to be
supported.</p>
</div>
</div>
</div>
</div>
</div>
<div class="section" id="line-number-information">
<span id="amdgpu-dwarf-line-number-information"></span><h4><a class="toc-backref" href="#id39">Line Number Information</a><a class="headerlink" href="#line-number-information" title="Permalink to this headline">¶</a></h4>
<div class="section" id="the-line-number-program-header">
<h5><a class="toc-backref" href="#id40">The Line Number Program Header</a><a class="headerlink" href="#the-line-number-program-header" title="Permalink to this headline">¶</a></h5>
<div class="section" id="standard-content-descriptions">
<h6><a class="toc-backref" href="#id41">Standard Content Descriptions</a><a class="headerlink" href="#standard-content-descriptions" title="Permalink to this headline">¶</a></h6>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 6.2.4.1.</p>
</div>
<ol class="arabic" id="amdgpu-dwarf-line-number-information-dw-lnct-llvm-source">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_LNCT_LLVM_source</span></code></p>
<p>The component is a null-terminated UTF-8 source text string with “<code class="docutils literal notranslate"><span class="pre">\n</span></code>” line endings. This content code is paired with the same forms as
<code class="docutils literal notranslate"><span class="pre">DW_LNCT_path</span></code>. It can be used for file name entries.</p>
<p>The value is an empty null-terminated string if no source is available. If
the source is available but is an empty file then the value is a
null-terminated single “<code class="docutils literal notranslate"><span class="pre">\n</span></code>“.</p>
<p><em>When the source field is present, consumers can use the embedded source
instead of attempting to discover the source on disk using the file path
provided by the</em> <code class="docutils literal notranslate"><span class="pre">DW_LNCT_path</span></code> <em>field. When the source field is absent,
consumers can access the file to get the source text.</em></p>
<p><em>This is particularly useful for programming languages that support runtime
compilation and runtime generation of source text. In these cases, the
source text does not reside in any permanent file. For example, the OpenCL
language [:ref:`OpenCL <amdgpu-dwarf-OpenCL>`] supports online compilation.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_LNCT_LLVM_is_MD5</span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">DW_LNCT_LLVM_is_MD5</span></code> indicates if the <code class="docutils literal notranslate"><span class="pre">DW_LNCT_MD5</span></code> content kind, if
present, is valid: when 0 it is not valid and when 1 it is valid. If
<code class="docutils literal notranslate"><span class="pre">DW_LNCT_LLVM_is_MD5</span></code> content kind is not present, and <code class="docutils literal notranslate"><span class="pre">DW_LNCT_MD5</span></code>
content kind is present, then the MD5 checksum is valid.</p>
<p><code class="docutils literal notranslate"><span class="pre">DW_LNCT_LLVM_is_MD5</span></code> is always paired with the <code class="docutils literal notranslate"><span class="pre">DW_FORM_udata</span></code> form.</p>
<p><em>This allows a compilation unit to have a mixture of files with and without
MD5 checksums. This can happen when multiple relocatable files are linked
together.</em></p>
</li>
</ol>
</div>
</div>
</div>
<div class="section" id="call-frame-information">
<span id="amdgpu-dwarf-call-frame-information"></span><h4><a class="toc-backref" href="#id42">Call Frame Information</a><a class="headerlink" href="#call-frame-information" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This section provides changes to existing call frame information and defines
instructions added by these extensions. Additional support is added for
address spaces. Register unwind DWARF expressions are generalized to allow any
location description, including those with composite and implicit location
descriptions.</p>
<p>These changes would be incorporated into the DWARF Version 5 section 6.1.</p>
</div>
<div class="section" id="structure-of-call-frame-information">
<span id="amdgpu-dwarf-structure-of-call-frame-information"></span><h5><a class="toc-backref" href="#id43">Structure of Call Frame Information</a><a class="headerlink" href="#structure-of-call-frame-information" title="Permalink to this headline">¶</a></h5>
<p>The register rules are:</p>
<dl>
<dt><em>undefined</em></dt><dd><p>A register that has this rule has no recoverable value in the previous frame.
The previous value of this register is the undefined location description (see
<a class="reference internal" href="#amdgpu-dwarf-undefined-location-description-operations"><span class="std std-ref">Undefined Location Description Operations</span></a>).</p>
<p><em>By convention, the register is not preserved by a callee.</em></p>
</dd>
<dt><em>same value</em></dt><dd><p>This register has not been modified from the previous caller frame.</p>
<p>If the current frame is the top frame, then the previous value of this
register is the location description L that specifies one register location
description SL. SL specifies the register location storage that corresponds to
the register with a bit offset of 0 for the current thread.</p>
<p>If the current frame is not the top frame, then the previous value of this
register is the location description obtained using the call frame information
for the callee frame and callee program location invoked by the current caller
frame for the same register.</p>
<p><em>By convention, the register is preserved by the callee, but the callee has
not modified it.</em></p>
</dd>
<dt><em>offset(N)</em></dt><dd><p>N is a signed byte offset. The previous value of this register is saved at the
location description computed as if the DWARF operation expression
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span> <span class="pre">N</span></code> is evaluated with the current context, except the
result kind is a location description, the compilation unit is unspecified,
the object is unspecified, and an initial stack comprising the location
description of the current CFA (see
<a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>).</p>
</dd>
<dt><em>val_offset(N)</em></dt><dd><p>N is a signed byte offset. The previous value of this register is the memory
byte address of the location description computed as if the DWARF operation
expression <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span> <span class="pre">N</span></code> is evaluated with the current context,
except the result kind is a location description, the compilation unit is
unspecified, the object is unspecified, and an initial stack comprising the
location description of the current CFA (see
<a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>).</p>
<p>The DWARF is ill-formed if the CFA location description is not a memory byte
address location description, or if the register size does not match the size
of an address in the address space of the current CFA location description.</p>
<p><em>Since the CFA location description is required to be a memory byte address
location description, the value of val_offset(N) will also be a memory byte
address location description since it is offsetting the CFA location
description by N bytes. Furthermore, the value of val_offset(N) will be a
memory byte address in the same address space as the CFA location
description.</em></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Should DWARF allow the address size to be a different size to the size of
the register? Requiring them to be the same bit size avoids any issue of
conversion as the bit contents of the register is simply interpreted as a
value of the address.</p>
<p>GDB has a per register hook that allows a target specific conversion on a
register by register basis. It defaults to truncation of bigger registers,
and to actually reading bytes from the next register (or reads out of bounds
for the last register) for smaller registers. There are no GDB tests that
read a register out of bounds (except an illegal hand written assembly
test).</p>
</div>
</dd>
<dt><em>register(R)</em></dt><dd><p>This register has been stored in another register numbered R.</p>
<p>The previous value of this register is the location description obtained using
the call frame information for the current frame and current program location
for register R.</p>
<p>The DWARF is ill-formed if the size of this register does not match the size
of register R or if there is a cyclic dependency in the call frame
information.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Should this also allow R to be larger than this register? If so is the value
stored in the low order bits and it is undefined what is stored in the
extra upper bits?</p>
</div>
</dd>
<dt><em>expression(E)</em></dt><dd><p>The previous value of this register is located at the location description
produced by evaluating the DWARF operation expression E (see
<a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>).</p>
<p>E is evaluated with the current context, except the result kind is a location
description, the compilation unit is unspecified, the object is unspecified,
and an initial stack comprising the location description of the current CFA
(see <a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>).</p>
</dd>
<dt><em>val_expression(E)</em></dt><dd><p>The previous value of this register is the value produced by evaluating the
DWARF operation expression E (see <a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>).</p>
<p>E is evaluated with the current context, except the result kind is a value,
the compilation unit is unspecified, the object is unspecified, and an initial
stack comprising the location description of the current CFA (see
<a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>).</p>
<p>The DWARF is ill-formed if the resulting value type size does not match the
register size.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This has limited usefulness as the DWARF expression E can only produce
values up to the size of the generic type. This is due to not allowing any
operations that specify a type in a CFI operation expression. This makes it
unusable for registers that are larger than the generic type. However,
<em>expression(E)</em> can be used to create an implicit location description of
any size.</p>
</div>
</dd>
<dt><em>architectural</em></dt><dd><p>The rule is defined externally to this specification by the augmenter.</p>
</dd>
</dl>
<p>A Common Information Entry (CIE) holds information that is shared among many
Frame Description Entries (FDE). There is at least one CIE in every non-empty
<code class="docutils literal notranslate"><span class="pre">.debug_frame</span></code> section. A CIE contains the following fields, in order:</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">length</span></code> (initial length)</p>
<p>A constant that gives the number of bytes of the CIE structure, not
including the length field itself. The size of the length field plus the
value of length must be an integral multiple of the address size specified
in the <code class="docutils literal notranslate"><span class="pre">address_size</span></code> field.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">CIE_id</span></code> (4 or 8 bytes, see
<a class="reference internal" href="#amdgpu-dwarf-32-bit-and-64-bit-dwarf-formats"><span class="std std-ref">32-Bit and 64-Bit DWARF Formats</span></a>)</p>
<p>A constant that is used to distinguish CIEs from FDEs.</p>
<p>In the 32-bit DWARF format, the value of the CIE id in the CIE header is
0xffffffff; in the 64-bit DWARF format, the value is 0xffffffffffffffff.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">version</span></code> (ubyte)</p>
<p>A version number. This number is specific to the call frame information and
is independent of the DWARF version number.</p>
<p>The value of the CIE version number is 4.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Would this be increased to 5 to reflect the changes in these extensions?</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">augmentation</span></code> (sequence of UTF-8 characters)</p>
<p>A null-terminated UTF-8 string that identifies the augmentation to this CIE
or to the FDEs that use it. If a reader encounters an augmentation string
that is unexpected, then only the following fields can be read:</p>
<ul class="simple">
<li><p>CIE: length, CIE_id, version, augmentation</p></li>
<li><p>FDE: length, CIE_pointer, initial_location, address_range</p></li>
</ul>
<p>If there is no augmentation, this value is a zero byte.</p>
<p><em>The augmentation string allows users to indicate that there is additional
vendor and target architecture specific information in the CIE or FDE which
is needed to virtually unwind a stack frame. For example, this might be
information about dynamically allocated data which needs to be freed on exit
from the routine.</em></p>
<p><em>Because the</em> <code class="docutils literal notranslate"><span class="pre">.debug_frame</span></code> <em>section is useful independently of any</em>
<code class="docutils literal notranslate"><span class="pre">.debug_info</span></code> <em>section, the augmentation string always uses UTF-8
encoding.</em></p>
<p>The recommended format for the augmentation string is:</p>
<blockquote>
<div><div class="line-block">
<div class="line"><code class="docutils literal notranslate"><span class="pre">[</span></code><em>vendor</em><code class="docutils literal notranslate"><span class="pre">:v</span></code><em>X</em><code class="docutils literal notranslate"><span class="pre">.</span></code><em>Y</em>[<code class="docutils literal notranslate"><span class="pre">:</span></code><em>options</em>]<code class="docutils literal notranslate"><span class="pre">]</span></code>*</div>
</div>
</div></blockquote>
<p>Where <em>vendor</em> is the producer, <code class="docutils literal notranslate"><span class="pre">vX.Y</span></code> specifies the major X and minor Y
version number of the extensions used, and <em>options</em> is an optional string
providing additional information about the extensions. The version number
must conform to semantic versioning [<a class="reference internal" href="#amdgpu-dwarf-semver"><span class="std std-ref">SEMVER</span></a>].
The <em>options</em> string must not contain the “<code class="docutils literal notranslate"><span class="pre">]</span></code>” character.</p>
<p>For example:</p>
<blockquote>
<div><div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="p">[</span><span class="n">abc</span><span class="p">:</span><span class="n">v0</span><span class="o">.</span><span class="mi">0</span><span class="p">][</span><span class="n">def</span><span class="p">:</span><span class="n">v1</span><span class="o">.</span><span class="mi">2</span><span class="p">:</span><span class="n">feature</span><span class="o">-</span><span class="n">a</span><span class="o">=</span><span class="n">on</span><span class="p">,</span><span class="n">feature</span><span class="o">-</span><span class="n">b</span><span class="o">=</span><span class="mi">3</span><span class="p">]</span>
</pre></div>
</div>
</div></blockquote>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">address_size</span></code> (ubyte)</p>
<p>The size of a target address in this CIE and any FDEs that use it, in bytes.
If a compilation unit exists for this frame, its address size must match the
address size here.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">segment_selector_size</span></code> (ubyte)</p>
<p>The size of a segment selector in this CIE and any FDEs that use it, in
bytes.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">code_alignment_factor</span></code> (unsigned LEB128)</p>
<p>A constant that is factored out of all advance location instructions (see
<a class="reference internal" href="#amdgpu-dwarf-row-creation-instructions"><span class="std std-ref">Row Creation Instructions</span></a>). The resulting value is
<code class="docutils literal notranslate"><span class="pre">(operand</span> <span class="pre">*</span> <span class="pre">code_alignment_factor)</span></code>.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">data_alignment_factor</span></code> (signed LEB128)</p>
<p>A constant that is factored out of certain offset instructions (see
<a class="reference internal" href="#amdgpu-dwarf-cfa-definition-instructions"><span class="std std-ref">CFA Definition Instructions</span></a> and
<a class="reference internal" href="#amdgpu-dwarf-register-rule-instructions"><span class="std std-ref">Register Rule Instructions</span></a>). The resulting value is
<code class="docutils literal notranslate"><span class="pre">(operand</span> <span class="pre">*</span> <span class="pre">data_alignment_factor)</span></code>.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">return_address_register</span></code> (unsigned LEB128)</p>
<p>An unsigned LEB128 constant that indicates which column in the rule table
represents the return address of the subprogram. Note that this column might
not correspond to an actual machine register.</p>
<p>The value of the return address register is used to determine the program
location of the caller frame. The program location of the top frame is the
target architecture program counter value of the current thread.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">initial_instructions</span></code> (array of ubyte)</p>
<p>A sequence of rules that are interpreted to create the initial setting of
each column in the table.</p>
<p>The default rule for all columns before interpretation of the initial
instructions is the undefined rule. However, an ABI authoring body or a
compilation system authoring body may specify an alternate default value for
any or all columns.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">padding</span></code> (array of ubyte)</p>
<p>Enough <code class="docutils literal notranslate"><span class="pre">DW_CFA_nop</span></code> instructions to make the size of this entry match the
length value above.</p>
</li>
</ol>
<p>An FDE contains the following fields, in order:</p>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">length</span></code> (initial length)</p>
<p>A constant that gives the number of bytes of the header and instruction
stream for this subprogram, not including the length field itself. The size
of the length field plus the value of length must be an integral multiple of
the address size.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">CIE_pointer</span></code> (4 or 8 bytes, see
<a class="reference internal" href="#amdgpu-dwarf-32-bit-and-64-bit-dwarf-formats"><span class="std std-ref">32-Bit and 64-Bit DWARF Formats</span></a>)</p>
<p>A constant offset into the <code class="docutils literal notranslate"><span class="pre">.debug_frame</span></code> section that denotes the CIE
that is associated with this FDE.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">initial_location</span></code> (segment selector and target address)</p>
<p>The address of the first location associated with this table entry. If the
segment_selector_size field of this FDE’s CIE is non-zero, the initial
location is preceded by a segment selector of the given length.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">address_range</span></code> (target address)</p>
<p>The number of bytes of program instructions described by this entry.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">instructions</span></code> (array of ubyte)</p>
<p>A sequence of table defining instructions that are described in
<a class="reference internal" href="#amdgpu-dwarf-call-frame-instructions"><span class="std std-ref">Call Frame Instructions</span></a>.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">padding</span></code> (array of ubyte)</p>
<p>Enough <code class="docutils literal notranslate"><span class="pre">DW_CFA_nop</span></code> instructions to make the size of this entry match the
length value above.</p>
</li>
</ol>
</div>
<div class="section" id="call-frame-instructions">
<span id="amdgpu-dwarf-call-frame-instructions"></span><h5><a class="toc-backref" href="#id44">Call Frame Instructions</a><a class="headerlink" href="#call-frame-instructions" title="Permalink to this headline">¶</a></h5>
<p>Some call frame instructions have operands that are encoded as DWARF operation
expressions E (see <a class="reference internal" href="#amdgpu-dwarf-operation-expressions"><span class="std std-ref">DWARF Operation Expressions</span></a>). The DWARF
operations that can be used in E have the following restrictions:</p>
<ul>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_addrx</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_call2</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_call4</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_call_ref</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_OP_const_type</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_constx</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_convert</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_OP_deref_type</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_fbreg</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_implicit_pointer</span></code>,
<code class="docutils literal notranslate"><span class="pre">DW_OP_regval_type</span></code>, <code class="docutils literal notranslate"><span class="pre">DW_OP_reinterpret</span></code>, and <code class="docutils literal notranslate"><span class="pre">DW_OP_xderef_type</span></code>
operations are not allowed because the call frame information must not depend
on other debug sections.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_push_object_address</span></code> is not allowed because there is no object
context to provide a value to push.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_push_lane</span></code> is not allowed because the call frame instructions
describe the actions for the whole thread, not the lanes independently.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_call_frame_cfa</span></code> and <code class="docutils literal notranslate"><span class="pre">DW_OP_entry_value</span></code> are not allowed because
their use would be circular.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_call_frame_entry_reg</span></code> is not allowed if evaluating E causes a
circular dependency between <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_call_frame_entry_reg</span></code> operations.</p>
<p><em>For example, if a register R1 has a</em> <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_expression</span></code>
<em>instruction that evaluates a</em> <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_call_frame_entry_reg</span></code> <em>operation
that specifies register R2, and register R2 has a</em>
<code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_expression</span></code> <em>instruction that that evaluates a</em>
<code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_call_frame_entry_reg</span></code> <em>operation that specifies register R1.</em></p>
</li>
</ul>
<p><em>Call frame instructions to which these restrictions apply include</em>
<code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_expression</span></code><em>,</em> <code class="docutils literal notranslate"><span class="pre">DW_CFA_expression</span></code><em>, and</em>
<code class="docutils literal notranslate"><span class="pre">DW_CFA_val_expression</span></code><em>.</em></p>
<div class="section" id="row-creation-instructions">
<span id="amdgpu-dwarf-row-creation-instructions"></span><h6><a class="toc-backref" href="#id45">Row Creation Instructions</a><a class="headerlink" href="#row-creation-instructions" title="Permalink to this headline">¶</a></h6>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>These instructions are the same as in DWARF Version 5 section 6.4.2.1.</p>
</div>
</div>
<div class="section" id="cfa-definition-instructions">
<span id="amdgpu-dwarf-cfa-definition-instructions"></span><h6><a class="toc-backref" href="#id46">CFA Definition Instructions</a><a class="headerlink" href="#cfa-definition-instructions" title="Permalink to this headline">¶</a></h6>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa</span></code> instruction takes two unsigned LEB128 operands
representing a register number R and a (non-factored) byte displacement B.
AS is set to the target architecture default address space identifier. The
required action is to define the current CFA rule to be the result of
evaluating the DWARF operation expression <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">AS;</span>
<span class="pre">DW_OP_aspace_bregx</span> <span class="pre">R,</span> <span class="pre">B</span></code> as a location description.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_sf</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_sf</span></code> instruction takes two operands: an unsigned LEB128
value representing a register number R and a signed LEB128 factored byte
displacement B. AS is set to the target architecture default address space
identifier. The required action is to define the current CFA rule to be the
result of evaluating the DWARF operation expression <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">AS;</span>
<span class="pre">DW_OP_aspace_bregx</span> <span class="pre">R,</span> <span class="pre">B*data_alignment_factor</span></code> as a location description.</p>
<p><em>The action is the same as</em> <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa</span></code><em>, except that the second
operand is signed and factored.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_LLVM_def_aspace_cfa</span></code> <em>New</em></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_LLVM_def_aspace_cfa</span></code> instruction takes three unsigned LEB128
operands representing a register number R, a (non-factored) byte
displacement B, and a target architecture specific address space identifier
AS. The required action is to define the current CFA rule to be the result
of evaluating the DWARF operation expression <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">AS;</span>
<span class="pre">DW_OP_aspace_bregx</span> <span class="pre">R,</span> <span class="pre">B</span></code> as a location description.</p>
<p>If AS is not one of the values defined by the target architecture specific
<code class="docutils literal notranslate"><span class="pre">DW_ASPACE_*</span></code> values then the DWARF expression is ill-formed.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_LLVM_def_aspace_cfa_sf</span></code> <em>New</em></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_sf</span></code> instruction takes three operands: an unsigned
LEB128 value representing a register number R, a signed LEB128 factored byte
displacement B, and an unsigned LEB128 value representing a target
architecture specific address space identifier AS. The required action is to
define the current CFA rule to be the result of evaluating the DWARF
operation expression <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">AS;</span> <span class="pre">DW_OP_aspace_bregx</span> <span class="pre">R,</span>
<span class="pre">B*data_alignment_factor</span></code> as a location description.</p>
<p>If AS is not one of the values defined by the target architecture specific
<code class="docutils literal notranslate"><span class="pre">DW_ASPACE_*</span></code> values, then the DWARF expression is ill-formed.</p>
<p><em>The action is the same as</em> <code class="docutils literal notranslate"><span class="pre">DW_CFA_aspace_def_cfa</span></code><em>, except that the
second operand is signed and factored.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_register</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_register</span></code> instruction takes a single unsigned LEB128
operand representing a register number R. The required action is to define
the current CFA rule to be the result of evaluating the DWARF operation
expression <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">AS;</span> <span class="pre">DW_OP_aspace_bregx</span> <span class="pre">R,</span> <span class="pre">B</span></code> as a location
description. B and AS are the old CFA byte displacement and address space
respectively.</p>
<p>If the subprogram has no current CFA rule, or the rule was defined by a
<code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_expression</span></code> instruction, then the DWARF is ill-formed.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_offset</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_offset</span></code> instruction takes a single unsigned LEB128
operand representing a (non-factored) byte displacement B. The required
action is to define the current CFA rule to be the result of evaluating the
DWARF operation expression <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">AS;</span> <span class="pre">DW_OP_aspace_bregx</span> <span class="pre">R,</span> <span class="pre">B</span></code> as a
location description. R and AS are the old CFA register number and address
space respectively.</p>
<p>If the subprogram has no current CFA rule, or the rule was defined by a
<code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_expression</span></code> instruction, then the DWARF is ill-formed.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_offset_sf</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_offset_sf</span></code> instruction takes a signed LEB128 operand
representing a factored byte displacement B. The required action is to
define the current CFA rule to be the result of evaluating the DWARF
operation expression <code class="docutils literal notranslate"><span class="pre">DW_OP_constu</span> <span class="pre">AS;</span> <span class="pre">DW_OP_aspace_bregx</span> <span class="pre">R,</span>
<span class="pre">B*data_alignment_factor</span></code> as a location description. R and AS are the old
CFA register number and address space respectively.</p>
<p>If the subprogram has no current CFA rule, or the rule was defined by a
<code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_expression</span></code> instruction, then the DWARF is ill-formed.</p>
<p><em>The action is the same as</em> <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_offset</span></code><em>, except that the
operand is signed and factored.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_expression</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_def_cfa_expression</span></code> instruction takes a single operand encoded
as a <code class="docutils literal notranslate"><span class="pre">DW_FORM_exprloc</span></code> value representing a DWARF operation expression E.
The required action is to define the current CFA rule to be the result of
evaluating E with the current context, except the result kind is a location
description, the compilation unit is unspecified, the object is unspecified,
and an empty initial stack.</p>
<p><em>See</em> <a class="reference internal" href="#amdgpu-dwarf-call-frame-instructions"><span class="std std-ref">Call Frame Instructions</span></a> <em>regarding restrictions on
the DWARF expression operations that can be used in E.</em></p>
<p>The DWARF is ill-formed if the result of evaluating E is not a memory byte
address location description.</p>
</li>
</ol>
</div>
<div class="section" id="register-rule-instructions">
<span id="amdgpu-dwarf-register-rule-instructions"></span><h6><a class="toc-backref" href="#id47">Register Rule Instructions</a><a class="headerlink" href="#register-rule-instructions" title="Permalink to this headline">¶</a></h6>
<ol class="arabic">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_undefined</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_undefined</span></code> instruction takes a single unsigned LEB128 operand
that represents a register number R. The required action is to set the rule
for the register specified by R to <code class="docutils literal notranslate"><span class="pre">undefined</span></code>.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_same_value</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_same_value</span></code> instruction takes a single unsigned LEB128 operand
that represents a register number R. The required action is to set the rule
for the register specified by R to <code class="docutils literal notranslate"><span class="pre">same</span> <span class="pre">value</span></code>.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_offset</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_offset</span></code> instruction takes two operands: a register number R
(encoded with the opcode) and an unsigned LEB128 constant representing a
factored displacement B. The required action is to change the rule for the
register specified by R to be an <em>offset(B*data_alignment_factor)</em> rule.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Seems this should be named <code class="docutils literal notranslate"><span class="pre">DW_CFA_offset_uf</span></code> since the offset is
unsigned factored.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_offset_extended</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_offset_extended</span></code> instruction takes two unsigned LEB128
operands representing a register number R and a factored displacement B.
This instruction is identical to <code class="docutils literal notranslate"><span class="pre">DW_CFA_offset</span></code>, except for the encoding
and size of the register operand.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Seems this should be named <code class="docutils literal notranslate"><span class="pre">DW_CFA_offset_extended_uf</span></code> since the
displacement is unsigned factored.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_offset_extended_sf</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_offset_extended_sf</span></code> instruction takes two operands: an
unsigned LEB128 value representing a register number R and a signed LEB128
factored displacement B. This instruction is identical to
<code class="docutils literal notranslate"><span class="pre">DW_CFA_offset_extended</span></code>, except that B is signed.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_val_offset</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_val_offset</span></code> instruction takes two unsigned LEB128 operands
representing a register number R and a factored displacement B. The required
action is to change the rule for the register indicated by R to be a
<em>val_offset(B*data_alignment_factor)</em> rule.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Seems this should be named <code class="docutils literal notranslate"><span class="pre">DW_CFA_val_offset_uf</span></code> since the displacement
is unsigned factored.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>An alternative is to define <code class="docutils literal notranslate"><span class="pre">DW_CFA_val_offset</span></code> to implicitly use the
target architecture default address space, and add another operation that
specifies the address space.</p>
</div>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_val_offset_sf</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_val_offset_sf</span></code> instruction takes two operands: an unsigned
LEB128 value representing a register number R and a signed LEB128 factored
displacement B. This instruction is identical to <code class="docutils literal notranslate"><span class="pre">DW_CFA_val_offset</span></code>,
except that B is signed.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_register</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_register</span></code> instruction takes two unsigned LEB128 operands
representing register numbers R1 and R2 respectively. The required action is
to set the rule for the register specified by R1 to be a <em>register(R2)</em> rule.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_expression</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_expression</span></code> instruction takes two operands: an unsigned LEB128
value representing a register number R, and a <code class="docutils literal notranslate"><span class="pre">DW_FORM_block</span></code> value
representing a DWARF operation expression E. The required action is to
change the rule for the register specified by R to be an <em>expression(E)</em>
rule.</p>
<p><em>That is, E computes the location description where the register value can
be retrieved.</em></p>
<p><em>See</em> <a class="reference internal" href="#amdgpu-dwarf-call-frame-instructions"><span class="std std-ref">Call Frame Instructions</span></a> <em>regarding restrictions on
the DWARF expression operations that can be used in E.</em></p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_val_expression</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_val_expression</span></code> instruction takes two operands: an unsigned
LEB128 value representing a register number R, and a <code class="docutils literal notranslate"><span class="pre">DW_FORM_block</span></code> value
representing a DWARF operation expression E. The required action is to
change the rule for the register specified by R to be a <em>val_expression(E)</em>
rule.</p>
<p><em>That is, E computes the value of register R.</em></p>
<p><em>See</em> <a class="reference internal" href="#amdgpu-dwarf-call-frame-instructions"><span class="std std-ref">Call Frame Instructions</span></a> <em>regarding restrictions on
the DWARF expression operations that can be used in E.</em></p>
<p>If the result of evaluating E is not a value with a base type size that
matches the register size, then the DWARF is ill-formed.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_restore</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_restore</span></code> instruction takes a single operand (encoded with the
opcode) that represents a register number R. The required action is to
change the rule for the register specified by R to the rule assigned it by
the <code class="docutils literal notranslate"><span class="pre">initial_instructions</span></code> in the CIE.</p>
</li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_CFA_restore_extended</span></code></p>
<p>The <code class="docutils literal notranslate"><span class="pre">DW_CFA_restore_extended</span></code> instruction takes a single unsigned LEB128
operand that represents a register number R. This instruction is identical
to <code class="docutils literal notranslate"><span class="pre">DW_CFA_restore</span></code>, except for the encoding and size of the register
operand.</p>
</li>
</ol>
</div>
<div class="section" id="row-state-instructions">
<h6><a class="toc-backref" href="#id48">Row State Instructions</a><a class="headerlink" href="#row-state-instructions" title="Permalink to this headline">¶</a></h6>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>These instructions are the same as in DWARF Version 5 section 6.4.2.4.</p>
</div>
</div>
<div class="section" id="padding-instruction">
<h6><a class="toc-backref" href="#id49">Padding Instruction</a><a class="headerlink" href="#padding-instruction" title="Permalink to this headline">¶</a></h6>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>These instructions are the same as in DWARF Version 5 section 6.4.2.5.</p>
</div>
</div>
</div>
<div class="section" id="call-frame-instruction-usage">
<h5><a class="toc-backref" href="#id50">Call Frame Instruction Usage</a><a class="headerlink" href="#call-frame-instruction-usage" title="Permalink to this headline">¶</a></h5>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The same as in DWARF Version 5 section 6.4.3.</p>
</div>
</div>
<div class="section" id="call-frame-calling-address">
<span id="amdgpu-dwarf-call-frame-calling-address"></span><h5><a class="toc-backref" href="#id51">Call Frame Calling Address</a><a class="headerlink" href="#call-frame-calling-address" title="Permalink to this headline">¶</a></h5>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The same as in DWARF Version 5 section 6.4.4.</p>
</div>
</div>
</div>
</div>
<div class="section" id="data-representation">
<h3><a class="toc-backref" href="#id52">Data Representation</a><a class="headerlink" href="#data-representation" title="Permalink to this headline">¶</a></h3>
<div class="section" id="bit-and-64-bit-dwarf-formats">
<span id="amdgpu-dwarf-32-bit-and-64-bit-dwarf-formats"></span><h4><a class="toc-backref" href="#id53">32-Bit and 64-Bit DWARF Formats</a><a class="headerlink" href="#bit-and-64-bit-dwarf-formats" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 7.4.</p>
</div>
<ol class="arabic">
<li><p>Within the body of the <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code> section, certain forms of attribute
value depend on the choice of DWARF format as follows. For the 32-bit DWARF
format, the value is a 4-byte unsigned integer; for the 64-bit DWARF format,
the value is an 8-byte unsigned integer.</p>
<table class="docutils align-default" id="amdgpu-dwarf-debug-info-section-attribute-form-roles-table">
<caption><span class="caption-text"><code class="docutils literal notranslate"><span class="pre">.debug_info</span></code> section attribute form roles</span><a class="headerlink" href="#amdgpu-dwarf-debug-info-section-attribute-form-roles-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 49%" />
<col style="width: 51%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Form</p></th>
<th class="head"><p>Role</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>DW_FORM_line_strp</p></td>
<td><p>offset in <code class="docutils literal notranslate"><span class="pre">.debug_line_str</span></code></p></td>
</tr>
<tr class="row-odd"><td><p>DW_FORM_ref_addr</p></td>
<td><p>offset in <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code></p></td>
</tr>
<tr class="row-even"><td><p>DW_FORM_sec_offset</p></td>
<td><p>offset in a section other than
<code class="docutils literal notranslate"><span class="pre">.debug_info</span></code> or <code class="docutils literal notranslate"><span class="pre">.debug_str</span></code></p></td>
</tr>
<tr class="row-odd"><td><p>DW_FORM_strp</p></td>
<td><p>offset in <code class="docutils literal notranslate"><span class="pre">.debug_str</span></code></p></td>
</tr>
<tr class="row-even"><td><p>DW_FORM_strp_sup</p></td>
<td><p>offset in <code class="docutils literal notranslate"><span class="pre">.debug_str</span></code> section of
supplementary object file</p></td>
</tr>
<tr class="row-odd"><td><p>DW_OP_call_ref</p></td>
<td><p>offset in <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code></p></td>
</tr>
<tr class="row-even"><td><p>DW_OP_implicit_pointer</p></td>
<td><p>offset in <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code></p></td>
</tr>
<tr class="row-odd"><td><p>DW_OP_LLVM_aspace_implicit_pointer</p></td>
<td><p>offset in <code class="docutils literal notranslate"><span class="pre">.debug_info</span></code></p></td>
</tr>
</tbody>
</table>
</li>
</ol>
</div>
<div class="section" id="format-of-debugging-information">
<h4><a class="toc-backref" href="#id54">Format of Debugging Information</a><a class="headerlink" href="#format-of-debugging-information" title="Permalink to this headline">¶</a></h4>
<div class="section" id="attribute-encodings">
<h5><a class="toc-backref" href="#id55">Attribute Encodings</a><a class="headerlink" href="#attribute-encodings" title="Permalink to this headline">¶</a></h5>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 7.5.4 and Table 7.5.</p>
</div>
<p>The following table gives the encoding of the additional debugging information
entry attributes.</p>
<table class="docutils align-default" id="amdgpu-dwarf-attribute-encodings-table">
<caption><span class="caption-text">Attribute encodings</span><a class="headerlink" href="#amdgpu-dwarf-attribute-encodings-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 45%" />
<col style="width: 8%" />
<col style="width: 47%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Attribute Name</p></th>
<th class="head"><p>Value</p></th>
<th class="head"><p>Classes</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>DW_AT_LLVM_active_lane</p></td>
<td><p>0x3e08</p></td>
<td><p>exprloc, loclist</p></td>
</tr>
<tr class="row-odd"><td><p>DW_AT_LLVM_augmentation</p></td>
<td><p>0x3e09</p></td>
<td><p>string</p></td>
</tr>
<tr class="row-even"><td><p>DW_AT_LLVM_lanes</p></td>
<td><p>0x3e0a</p></td>
<td><p>constant</p></td>
</tr>
<tr class="row-odd"><td><p>DW_AT_LLVM_lane_pc</p></td>
<td><p>0x3e0b</p></td>
<td><p>exprloc, loclist</p></td>
</tr>
<tr class="row-even"><td><p>DW_AT_LLVM_vector_size</p></td>
<td><p>0x3e0c</p></td>
<td><p>constant</p></td>
</tr>
</tbody>
</table>
</div>
</div>
<div class="section" id="id1">
<h4><a class="toc-backref" href="#id56">DWARF Expressions</a><a class="headerlink" href="#id1" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Rename DWARF Version 5 section 7.7 to reflect the unification of location
descriptions into DWARF expressions.</p>
</div>
<div class="section" id="operation-expressions">
<h5><a class="toc-backref" href="#id57">Operation Expressions</a><a class="headerlink" href="#operation-expressions" title="Permalink to this headline">¶</a></h5>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Rename DWARF Version 5 section 7.7.1 and delete section 7.7.2 to reflect the
unification of location descriptions into DWARF expressions.</p>
<p>This augments DWARF Version 5 section 7.7.1 and Table 7.9.</p>
</div>
<p>The following table gives the encoding of the additional DWARF expression
operations.</p>
<table class="docutils align-default" id="amdgpu-dwarf-operation-encodings-table">
<caption><span class="caption-text">DWARF Operation Encodings</span><a class="headerlink" href="#amdgpu-dwarf-operation-encodings-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 44%" />
<col style="width: 6%" />
<col style="width: 10%" />
<col style="width: 40%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Operation</p></th>
<th class="head"><p>Code</p></th>
<th class="head"><p>Number
of
Operands</p></th>
<th class="head"><p>Notes</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>DW_OP_LLVM_form_aspace_address</p></td>
<td><p>0xe1</p></td>
<td><p>0</p></td>
<td></td>
</tr>
<tr class="row-odd"><td><p>DW_OP_LLVM_push_lane</p></td>
<td><p>0xe2</p></td>
<td><p>0</p></td>
<td></td>
</tr>
<tr class="row-even"><td><p>DW_OP_LLVM_offset</p></td>
<td><p>0xe3</p></td>
<td><p>0</p></td>
<td></td>
</tr>
<tr class="row-odd"><td><p>DW_OP_LLVM_offset_uconst</p></td>
<td><p>0xe4</p></td>
<td><p>1</p></td>
<td><p>ULEB128 byte displacement</p></td>
</tr>
<tr class="row-even"><td><p>DW_OP_LLVM_bit_offset</p></td>
<td><p>0xe5</p></td>
<td><p>0</p></td>
<td></td>
</tr>
<tr class="row-odd"><td><p>DW_OP_LLVM_call_frame_entry_reg</p></td>
<td><p>0xe6</p></td>
<td><p>1</p></td>
<td><p>ULEB128 register number</p></td>
</tr>
<tr class="row-even"><td><p>DW_OP_LLVM_undefined</p></td>
<td><p>0xe7</p></td>
<td><p>0</p></td>
<td></td>
</tr>
<tr class="row-odd"><td><p>DW_OP_LLVM_aspace_bregx</p></td>
<td><p>0xe8</p></td>
<td><p>2</p></td>
<td><p>ULEB128 register number,
ULEB128 byte displacement</p></td>
</tr>
<tr class="row-even"><td><p>DW_OP_LLVM_aspace_implicit_pointer</p></td>
<td><p>0xe9</p></td>
<td><p>2</p></td>
<td><p>4-byte or 8-byte offset of DIE,
SLEB128 byte displacement</p></td>
</tr>
<tr class="row-odd"><td><p>DW_OP_LLVM_piece_end</p></td>
<td><p>0xea</p></td>
<td><p>0</p></td>
<td></td>
</tr>
<tr class="row-even"><td><p>DW_OP_LLVM_extend</p></td>
<td><p>0xeb</p></td>
<td><p>2</p></td>
<td><p>ULEB128 bit size,
ULEB128 count</p></td>
</tr>
<tr class="row-odd"><td><p>DW_OP_LLVM_select_bit_piece</p></td>
<td><p>0xec</p></td>
<td><p>2</p></td>
<td><p>ULEB128 bit size,
ULEB128 count</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="location-list-expressions">
<h5><a class="toc-backref" href="#id58">Location List Expressions</a><a class="headerlink" href="#location-list-expressions" title="Permalink to this headline">¶</a></h5>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Rename DWARF Version 5 section 7.7.3 to reflect that location lists are a kind
of DWARF expression.</p>
</div>
</div>
</div>
<div class="section" id="source-languages">
<h4><a class="toc-backref" href="#id59">Source Languages</a><a class="headerlink" href="#source-languages" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 7.12 and Table 7.17.</p>
</div>
<p>The following table gives the encoding of the additional DWARF languages.</p>
<table class="docutils align-default" id="amdgpu-dwarf-language-encodings-table">
<caption><span class="caption-text">Language encodings</span><a class="headerlink" href="#amdgpu-dwarf-language-encodings-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 44%" />
<col style="width: 13%" />
<col style="width: 42%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Language Name</p></th>
<th class="head"><p>Value</p></th>
<th class="head"><p>Default Lower Bound</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_LANG_LLVM_HIP</span></code></p></td>
<td><p>0x8100</p></td>
<td><p>0</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="address-class-and-address-space-encodings">
<h4><a class="toc-backref" href="#id60">Address Class and Address Space Encodings</a><a class="headerlink" href="#address-class-and-address-space-encodings" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This replaces DWARF Version 5 section 7.13.</p>
</div>
<p>The encodings of the constants used for the currently defined address classes
are given in <a class="reference internal" href="#amdgpu-dwarf-address-class-encodings-table"><span class="std std-ref">Address class encodings</span></a>.</p>
<table class="docutils align-default" id="amdgpu-dwarf-address-class-encodings-table">
<caption><span class="caption-text">Address class encodings</span><a class="headerlink" href="#amdgpu-dwarf-address-class-encodings-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 81%" />
<col style="width: 19%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Address Class Name</p></th>
<th class="head"><p>Value</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_none</span></code></p></td>
<td><p>0x0000</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_global</span></code></p></td>
<td><p>0x0001</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_constant</span></code></p></td>
<td><p>0x0002</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_group</span></code></p></td>
<td><p>0x0003</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_private</span></code></p></td>
<td><p>0x0004</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_lo_user</span></code></p></td>
<td><p>0x8000</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_ADDR_LLVM_hi_user</span></code></p></td>
<td><p>0xffff</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="id2">
<h4><a class="toc-backref" href="#id61">Line Number Information</a><a class="headerlink" href="#id2" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 7.22 and Table 7.27.</p>
</div>
<p>The following table gives the encoding of the additional line number header
entry formats.</p>
<table class="docutils align-default" id="amdgpu-dwarf-line-number-header-entry-format-encodings-table">
<caption><span class="caption-text">Line number header entry format encodings</span><a class="headerlink" href="#amdgpu-dwarf-line-number-header-entry-format-encodings-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 64%" />
<col style="width: 36%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Line number header entry format name</p></th>
<th class="head"><p>Value</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_LNCT_LLVM_source</span></code></p></td>
<td><p>0x2001</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_LNCT_LLVM_is_MD5</span></code></p></td>
<td><p>0x2002</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="id3">
<h4><a class="toc-backref" href="#id62">Call Frame Information</a><a class="headerlink" href="#id3" title="Permalink to this headline">¶</a></h4>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 section 7.24 and Table 7.29.</p>
</div>
<p>The following table gives the encoding of the additional call frame information
instructions.</p>
<table class="docutils align-default" id="amdgpu-dwarf-call-frame-instruction-encodings-table">
<caption><span class="caption-text">Call frame instruction encodings</span><a class="headerlink" href="#amdgpu-dwarf-call-frame-instruction-encodings-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 31%" />
<col style="width: 6%" />
<col style="width: 6%" />
<col style="width: 17%" />
<col style="width: 17%" />
<col style="width: 22%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Instruction</p></th>
<th class="head"><p>High 2
Bits</p></th>
<th class="head"><p>Low 6
Bits</p></th>
<th class="head"><p>Operand 1</p></th>
<th class="head"><p>Operand 2</p></th>
<th class="head"><p>Operand 3</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>DW_CFA_LLVM_def_aspace_cfa</p></td>
<td><p>0</p></td>
<td><p>0x30</p></td>
<td><p>ULEB128 register</p></td>
<td><p>ULEB128 offset</p></td>
<td><p>ULEB128 address space</p></td>
</tr>
<tr class="row-odd"><td><p>DW_CFA_LLVM_def_aspace_cfa_sf</p></td>
<td><p>0</p></td>
<td><p>0x31</p></td>
<td><p>ULEB128 register</p></td>
<td><p>SLEB128 offset</p></td>
<td><p>ULEB128 address space</p></td>
</tr>
</tbody>
</table>
</div>
</div>
<div class="section" id="attributes-by-tag-value-informative">
<h3><a class="toc-backref" href="#id63">Attributes by Tag Value (Informative)</a><a class="headerlink" href="#attributes-by-tag-value-informative" title="Permalink to this headline">¶</a></h3>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This augments DWARF Version 5 Appendix A and Table A.1.</p>
</div>
<p>The following table provides the additional attributes that are applicable to
debugger information entries.</p>
<table class="docutils align-default" id="amdgpu-dwarf-attributes-by-tag-value-table">
<caption><span class="caption-text">Attributes by tag value</span><a class="headerlink" href="#amdgpu-dwarf-attributes-by-tag-value-table" title="Permalink to this table">¶</a></caption>
<colgroup>
<col style="width: 50%" />
<col style="width: 50%" />
</colgroup>
<thead>
<tr class="row-odd"><th class="head"><p>Tag Name</p></th>
<th class="head"><p>Applicable Attributes</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_TAG_base_type</span></code></p></td>
<td><ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_vector_size</span></code></p></li>
</ul>
</td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_TAG_compile_unit</span></code></p></td>
<td><ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_augmentation</span></code></p></li>
</ul>
</td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_TAG_entry_point</span></code></p></td>
<td><ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_active_lane</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lanes</span></code></p></li>
</ul>
</td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">DW_TAG_inlined_subroutine</span></code></p></td>
<td><ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_active_lane</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lanes</span></code></p></li>
</ul>
</td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">DW_TAG_subprogram</span></code></p></td>
<td><ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_active_lane</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lane_pc</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">DW_AT_LLVM_lanes</span></code></p></li>
</ul>
</td>
</tr>
</tbody>
</table>
</div>
</div>
<div class="section" id="examples">
<span id="amdgpu-dwarf-examples"></span><h2><a class="toc-backref" href="#id64">Examples</a><a class="headerlink" href="#examples" title="Permalink to this headline">¶</a></h2>
<p>The AMD GPU specific usage of the features in these extensions, including
examples, is available at <em>User Guide for AMDGPU Backend</em> section
<a class="reference internal" href="AMDGPUUsage.html#amdgpu-dwarf-debug-information"><span class="std std-ref">DWARF Debug Information</span></a>.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Change examples to use <code class="docutils literal notranslate"><span class="pre">DW_OP_LLVM_offset</span></code> instead of <code class="docutils literal notranslate"><span class="pre">DW_OP_add</span></code> when
acting on a location description.</p>
<p>Need to provide examples of new features.</p>
</div>
</div>
<div class="section" id="references">
<span id="amdgpu-dwarf-references"></span><h2><a class="toc-backref" href="#id65">References</a><a class="headerlink" href="#references" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div></div></blockquote>
<ol class="arabic simple" id="amdgpu-dwarf-amd">
<li><p>[AMD] <a class="reference external" href="https://www.amd.com/">Advanced Micro Devices</a></p>
</li>
<li id="amdgpu-dwarf-amd-rocm"><p>[AMD-ROCm] <a class="reference external" href="https://rocm-documentation.readthedocs.io">AMD ROCm Platform</a></p>
</li>
<li id="amdgpu-dwarf-amd-rocgdb"><p>[AMD-ROCgdb] <a class="reference external" href="https://github.com/ROCm-Developer-Tools/ROCgdb">AMD ROCm Debugger (ROCgdb)</a></p>
</li>
<li id="amdgpu-dwarf-amdgpu-llvm"><p>[AMDGPU-LLVM] <a class="reference external" href="https://llvm.org/docs/AMDGPUUsage.html">User Guide for AMDGPU LLVM Backend</a></p>
</li>
<li id="amdgpu-dwarf-cuda"><p>[CUDA] <a class="reference external" href="https://docs.nvidia.com/cuda/cuda-c-programming-guide/">Nvidia CUDA Language</a></p>
</li>
<li id="amdgpu-dwarf-dwarf"><p>[DWARF] <a class="reference external" href="http://dwarfstd.org/">DWARF Debugging Information Format</a></p>
</li>
<li id="amdgpu-dwarf-elf"><p>[ELF] <a class="reference external" href="http://www.sco.com/developers/gabi/">Executable and Linkable Format (ELF)</a></p>
</li>
<li id="amdgpu-dwarf-gcc"><p>[GCC] <a class="reference external" href="https://www.gnu.org/software/gcc/">GCC: The GNU Compiler Collection</a></p>
</li>
<li id="amdgpu-dwarf-gdb"><p>[GDB] <a class="reference external" href="https://www.gnu.org/software/gdb/">GDB: The GNU Project Debugger</a></p>
</li>
<li id="amdgpu-dwarf-hip"><p>[HIP] <a class="reference external" href="https://rocm-documentation.readthedocs.io/en/latest/Programming_Guides/Programming-Guides.html#hip-programing-guide">HIP Programming Guide</a></p>
</li>
<li id="amdgpu-dwarf-hsa"><p>[HSA] <a class="reference external" href="http://www.hsafoundation.com/">Heterogeneous System Architecture (HSA) Foundation</a></p>
</li>
<li id="amdgpu-dwarf-llvm"><p>[LLVM] <a class="reference external" href="https://llvm.org/">The LLVM Compiler Infrastructure</a></p>
</li>
<li id="amdgpu-dwarf-opencl"><p>[OpenCL] <a class="reference external" href="http://www.khronos.org/registry/cl/specs/opencl-2.0.pdf">The OpenCL Specification Version 2.0</a></p>
</li>
<li id="amdgpu-dwarf-perforce-totalview"><p>[Perforce-TotalView] <a class="reference external" href="https://totalview.io/products/totalview">Perforce TotalView HPC Debugging Software</a></p>
</li>
<li id="amdgpu-dwarf-semver"><p>[SEMVER] <a class="reference external" href="https://semver.org/">Semantic Versioning</a></p></li>
</ol>
</div>
</div>
<div class="clearer"></div>
</div>
</div>
</div>
<div class="clearer"></div>
</div>
<div class="related" role="navigation" aria-label="related navigation">
<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="Benchmarking.html" title="Benchmarking tips"
>next</a> |</li>
<li class="right" >
<a href="AMDGPUInstructionNotation.html" title="AMDGPU Instructions Notation"
>previous</a> |</li>
<li><a href="https://llvm.org/">LLVM Home</a> | </li>
<li><a href="index.html">Documentation</a>»</li>
<li class="nav-item nav-item-1"><a href="UserGuides.html" >User Guides</a> »</li>
<li class="nav-item nav-item-2"><a href="AMDGPUUsage.html" >User Guide for AMDGPU Backend</a> »</li>
<li class="nav-item nav-item-this"><a href="">DWARF Extensions For Heterogeneous Debugging</a></li>
</ul>
</div>
<div class="footer" role="contentinfo">
© Copyright 2003-2021, LLVM Project.
Last updated on 2021-09-18.
Created using <a href="https://www.sphinx-doc.org/">Sphinx</a> 3.5.4.
</div>
</body>
</html>
|
