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
|
// dwarf_reader.cc -- parse dwarf2/3 debug information
// Copyright 2007, 2008, 2009, 2010, 2011, 2012 Free Software Foundation, Inc.
// Written by Ian Lance Taylor <iant@google.com>.
// This file is part of gold.
// This program is free software; you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation; either version 3 of the License, or
// (at your option) any later version.
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
// You should have received a copy of the GNU General Public License
// along with this program; if not, write to the Free Software
// Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
// MA 02110-1301, USA.
#include "gold.h"
#include <algorithm>
#include <vector>
#include "elfcpp_swap.h"
#include "dwarf.h"
#include "object.h"
#include "reloc.h"
#include "dwarf_reader.h"
#include "int_encoding.h"
#include "compressed_output.h"
namespace gold {
// Class Sized_elf_reloc_mapper
// Initialize the relocation tracker for section RELOC_SHNDX.
template<int size, bool big_endian>
bool
Sized_elf_reloc_mapper<size, big_endian>::do_initialize(
unsigned int reloc_shndx, unsigned int reloc_type)
{
this->reloc_type_ = reloc_type;
return this->track_relocs_.initialize(this->object_, reloc_shndx,
reloc_type);
}
// Looks in the symtab to see what section a symbol is in.
template<int size, bool big_endian>
unsigned int
Sized_elf_reloc_mapper<size, big_endian>::symbol_section(
unsigned int symndx, Address* value, bool* is_ordinary)
{
const int symsize = elfcpp::Elf_sizes<size>::sym_size;
gold_assert(static_cast<off_t>((symndx + 1) * symsize) <= this->symtab_size_);
elfcpp::Sym<size, big_endian> elfsym(this->symtab_ + symndx * symsize);
*value = elfsym.get_st_value();
return this->object_->adjust_sym_shndx(symndx, elfsym.get_st_shndx(),
is_ordinary);
}
// Return the section index and offset within the section of
// the target of the relocation for RELOC_OFFSET.
template<int size, bool big_endian>
unsigned int
Sized_elf_reloc_mapper<size, big_endian>::do_get_reloc_target(
off_t reloc_offset, off_t* target_offset)
{
this->track_relocs_.advance(reloc_offset);
if (reloc_offset != this->track_relocs_.next_offset())
return 0;
unsigned int symndx = this->track_relocs_.next_symndx();
typename elfcpp::Elf_types<size>::Elf_Addr value;
bool is_ordinary;
unsigned int target_shndx = this->symbol_section(symndx, &value,
&is_ordinary);
if (!is_ordinary)
return 0;
if (this->reloc_type_ == elfcpp::SHT_RELA)
value += this->track_relocs_.next_addend();
*target_offset = value;
return target_shndx;
}
static inline Elf_reloc_mapper*
make_elf_reloc_mapper(Relobj* object, const unsigned char* symtab,
off_t symtab_size)
{
if (object->elfsize() == 32)
{
if (object->is_big_endian())
{
#ifdef HAVE_TARGET_32_BIG
return new Sized_elf_reloc_mapper<32, true>(object, symtab,
symtab_size);
#else
gold_unreachable();
#endif
}
else
{
#ifdef HAVE_TARGET_32_LITTLE
return new Sized_elf_reloc_mapper<32, false>(object, symtab,
symtab_size);
#else
gold_unreachable();
#endif
}
}
else if (object->elfsize() == 64)
{
if (object->is_big_endian())
{
#ifdef HAVE_TARGET_64_BIG
return new Sized_elf_reloc_mapper<64, true>(object, symtab,
symtab_size);
#else
gold_unreachable();
#endif
}
else
{
#ifdef HAVE_TARGET_64_LITTLE
return new Sized_elf_reloc_mapper<64, false>(object, symtab,
symtab_size);
#else
gold_unreachable();
#endif
}
}
else
gold_unreachable();
}
// class Dwarf_abbrev_table
void
Dwarf_abbrev_table::clear_abbrev_codes()
{
for (unsigned int code = 0; code < this->low_abbrev_code_max_; ++code)
{
if (this->low_abbrev_codes_[code] != NULL)
{
delete this->low_abbrev_codes_[code];
this->low_abbrev_codes_[code] = NULL;
}
}
for (Abbrev_code_table::iterator it = this->high_abbrev_codes_.begin();
it != this->high_abbrev_codes_.end();
++it)
{
if (it->second != NULL)
delete it->second;
}
this->high_abbrev_codes_.clear();
}
// Read the abbrev table from an object file.
bool
Dwarf_abbrev_table::do_read_abbrevs(
Relobj* object,
unsigned int abbrev_shndx,
off_t abbrev_offset)
{
this->clear_abbrev_codes();
// If we don't have relocations, abbrev_shndx will be 0, and
// we'll have to hunt for the .debug_abbrev section.
if (abbrev_shndx == 0 && this->abbrev_shndx_ > 0)
abbrev_shndx = this->abbrev_shndx_;
else if (abbrev_shndx == 0)
{
for (unsigned int i = 1; i < object->shnum(); ++i)
{
std::string name = object->section_name(i);
if (name == ".debug_abbrev")
{
abbrev_shndx = i;
// Correct the offset. For incremental update links, we have a
// relocated offset that is relative to the output section, but
// here we need an offset relative to the input section.
abbrev_offset -= object->output_section_offset(i);
break;
}
}
if (abbrev_shndx == 0)
return false;
}
// Get the section contents and decompress if necessary.
if (abbrev_shndx != this->abbrev_shndx_)
{
if (this->owns_buffer_ && this->buffer_ != NULL)
{
delete[] this->buffer_;
this->owns_buffer_ = false;
}
section_size_type buffer_size;
this->buffer_ =
object->decompressed_section_contents(abbrev_shndx,
&buffer_size,
&this->owns_buffer_);
this->buffer_end_ = this->buffer_ + buffer_size;
this->abbrev_shndx_ = abbrev_shndx;
}
this->buffer_pos_ = this->buffer_ + abbrev_offset;
return true;
}
// Lookup the abbrev code entry for CODE. This function is called
// only when the abbrev code is not in the direct lookup table.
// It may be in the hash table, it may not have been read yet,
// or it may not exist in the abbrev table.
const Dwarf_abbrev_table::Abbrev_code*
Dwarf_abbrev_table::do_get_abbrev(unsigned int code)
{
// See if the abbrev code is already in the hash table.
Abbrev_code_table::const_iterator it = this->high_abbrev_codes_.find(code);
if (it != this->high_abbrev_codes_.end())
return it->second;
// Read and store abbrev code definitions until we find the
// one we're looking for.
for (;;)
{
// Read the abbrev code. A zero here indicates the end of the
// abbrev table.
size_t len;
if (this->buffer_pos_ >= this->buffer_end_)
return NULL;
uint64_t nextcode = read_unsigned_LEB_128(this->buffer_pos_, &len);
if (nextcode == 0)
{
this->buffer_pos_ = this->buffer_end_;
return NULL;
}
this->buffer_pos_ += len;
// Read the tag.
if (this->buffer_pos_ >= this->buffer_end_)
return NULL;
uint64_t tag = read_unsigned_LEB_128(this->buffer_pos_, &len);
this->buffer_pos_ += len;
// Read the has_children flag.
if (this->buffer_pos_ >= this->buffer_end_)
return NULL;
bool has_children = *this->buffer_pos_ == elfcpp::DW_CHILDREN_yes;
this->buffer_pos_ += 1;
// Read the list of (attribute, form) pairs.
Abbrev_code* entry = new Abbrev_code(tag, has_children);
for (;;)
{
// Read the attribute.
if (this->buffer_pos_ >= this->buffer_end_)
return NULL;
uint64_t attr = read_unsigned_LEB_128(this->buffer_pos_, &len);
this->buffer_pos_ += len;
// Read the form.
if (this->buffer_pos_ >= this->buffer_end_)
return NULL;
uint64_t form = read_unsigned_LEB_128(this->buffer_pos_, &len);
this->buffer_pos_ += len;
// A (0,0) pair terminates the list.
if (attr == 0 && form == 0)
break;
if (attr == elfcpp::DW_AT_sibling)
entry->has_sibling_attribute = true;
entry->add_attribute(attr, form);
}
this->store_abbrev(nextcode, entry);
if (nextcode == code)
return entry;
}
return NULL;
}
// class Dwarf_ranges_table
// Read the ranges table from an object file.
bool
Dwarf_ranges_table::read_ranges_table(
Relobj* object,
const unsigned char* symtab,
off_t symtab_size,
unsigned int ranges_shndx)
{
// If we've already read this abbrev table, return immediately.
if (this->ranges_shndx_ > 0
&& this->ranges_shndx_ == ranges_shndx)
return true;
// If we don't have relocations, ranges_shndx will be 0, and
// we'll have to hunt for the .debug_ranges section.
if (ranges_shndx == 0 && this->ranges_shndx_ > 0)
ranges_shndx = this->ranges_shndx_;
else if (ranges_shndx == 0)
{
for (unsigned int i = 1; i < object->shnum(); ++i)
{
std::string name = object->section_name(i);
if (name == ".debug_ranges")
{
ranges_shndx = i;
this->output_section_offset_ = object->output_section_offset(i);
break;
}
}
if (ranges_shndx == 0)
return false;
}
// Get the section contents and decompress if necessary.
if (ranges_shndx != this->ranges_shndx_)
{
if (this->owns_ranges_buffer_ && this->ranges_buffer_ != NULL)
{
delete[] this->ranges_buffer_;
this->owns_ranges_buffer_ = false;
}
section_size_type buffer_size;
this->ranges_buffer_ =
object->decompressed_section_contents(ranges_shndx,
&buffer_size,
&this->owns_ranges_buffer_);
this->ranges_buffer_end_ = this->ranges_buffer_ + buffer_size;
this->ranges_shndx_ = ranges_shndx;
}
if (this->ranges_reloc_mapper_ != NULL)
{
delete this->ranges_reloc_mapper_;
this->ranges_reloc_mapper_ = NULL;
}
// For incremental objects, we have no relocations.
if (object->is_incremental())
return true;
// Find the relocation section for ".debug_ranges".
unsigned int reloc_shndx = 0;
unsigned int reloc_type = 0;
for (unsigned int i = 0; i < object->shnum(); ++i)
{
reloc_type = object->section_type(i);
if ((reloc_type == elfcpp::SHT_REL
|| reloc_type == elfcpp::SHT_RELA)
&& object->section_info(i) == ranges_shndx)
{
reloc_shndx = i;
break;
}
}
this->ranges_reloc_mapper_ = make_elf_reloc_mapper(object, symtab,
symtab_size);
this->ranges_reloc_mapper_->initialize(reloc_shndx, reloc_type);
return true;
}
// Read a range list from section RANGES_SHNDX at offset RANGES_OFFSET.
Dwarf_range_list*
Dwarf_ranges_table::read_range_list(
Relobj* object,
const unsigned char* symtab,
off_t symtab_size,
unsigned int addr_size,
unsigned int ranges_shndx,
off_t offset)
{
Dwarf_range_list* ranges;
if (!this->read_ranges_table(object, symtab, symtab_size, ranges_shndx))
return NULL;
// Correct the offset. For incremental update links, we have a
// relocated offset that is relative to the output section, but
// here we need an offset relative to the input section.
offset -= this->output_section_offset_;
// Read the range list at OFFSET.
ranges = new Dwarf_range_list();
off_t base = 0;
for (;
this->ranges_buffer_ + offset < this->ranges_buffer_end_;
offset += 2 * addr_size)
{
off_t start;
off_t end;
// Read the raw contents of the section.
if (addr_size == 4)
{
start = this->dwinfo_->read_from_pointer<32>(this->ranges_buffer_
+ offset);
end = this->dwinfo_->read_from_pointer<32>(this->ranges_buffer_
+ offset + 4);
}
else
{
start = this->dwinfo_->read_from_pointer<64>(this->ranges_buffer_
+ offset);
end = this->dwinfo_->read_from_pointer<64>(this->ranges_buffer_
+ offset + 8);
}
// Check for relocations and adjust the values.
unsigned int shndx1 = 0;
unsigned int shndx2 = 0;
if (this->ranges_reloc_mapper_ != NULL)
{
shndx1 =
this->ranges_reloc_mapper_->get_reloc_target(offset, &start);
shndx2 =
this->ranges_reloc_mapper_->get_reloc_target(offset + addr_size,
&end);
}
// End of list is marked by a pair of zeroes.
if (shndx1 == 0 && start == 0 && end == 0)
break;
// A "base address selection entry" is identified by
// 0xffffffff for the first value of the pair. The second
// value is used as a base for subsequent range list entries.
if (shndx1 == 0 && start == -1)
base = end;
else if (shndx1 == shndx2)
{
if (shndx1 == 0 || object->is_section_included(shndx1))
ranges->add(shndx1, base + start, base + end);
}
else
gold_warning(_("%s: DWARF info may be corrupt; offsets in a "
"range list entry are in different sections"),
object->name().c_str());
}
return ranges;
}
// class Dwarf_pubnames_table
// Read the pubnames section SHNDX from the object file.
bool
Dwarf_pubnames_table::read_section(Relobj* object, unsigned int shndx)
{
section_size_type buffer_size;
// If we don't have relocations, shndx will be 0, and
// we'll have to hunt for the .debug_pubnames/pubtypes section.
if (shndx == 0)
{
const char* name = (this->is_pubtypes_
? ".debug_pubtypes"
: ".debug_pubnames");
for (unsigned int i = 1; i < object->shnum(); ++i)
{
if (object->section_name(i) == name)
{
shndx = i;
this->output_section_offset_ = object->output_section_offset(i);
break;
}
}
if (shndx == 0)
return false;
}
this->buffer_ = object->decompressed_section_contents(shndx,
&buffer_size,
&this->owns_buffer_);
if (this->buffer_ == NULL)
return false;
this->buffer_end_ = this->buffer_ + buffer_size;
return true;
}
// Read the header for the set at OFFSET.
bool
Dwarf_pubnames_table::read_header(off_t offset)
{
// Correct the offset. For incremental update links, we have a
// relocated offset that is relative to the output section, but
// here we need an offset relative to the input section.
offset -= this->output_section_offset_;
if (offset < 0 || offset + 14 >= this->buffer_end_ - this->buffer_)
return false;
const unsigned char* pinfo = this->buffer_ + offset;
// Read the unit_length field.
uint32_t unit_length = this->dwinfo_->read_from_pointer<32>(pinfo);
pinfo += 4;
if (unit_length == 0xffffffff)
{
unit_length = this->dwinfo_->read_from_pointer<64>(pinfo);
pinfo += 8;
this->offset_size_ = 8;
}
else
this->offset_size_ = 4;
// Check the version.
unsigned int version = this->dwinfo_->read_from_pointer<16>(pinfo);
pinfo += 2;
if (version != 2)
return false;
// Skip the debug_info_offset and debug_info_size fields.
pinfo += 2 * this->offset_size_;
if (pinfo >= this->buffer_end_)
return false;
this->pinfo_ = pinfo;
return true;
}
// Read the next name from the set.
const char*
Dwarf_pubnames_table::next_name()
{
const unsigned char* pinfo = this->pinfo_;
// Read the offset within the CU. If this is zero, we have reached
// the end of the list.
uint32_t offset;
if (this->offset_size_ == 4)
offset = this->dwinfo_->read_from_pointer<32>(&pinfo);
else
offset = this->dwinfo_->read_from_pointer<64>(&pinfo);
if (offset == 0)
return NULL;
// Return a pointer to the string at the current location,
// and advance the pointer to the next entry.
const char* ret = reinterpret_cast<const char*>(pinfo);
while (pinfo < this->buffer_end_ && *pinfo != '\0')
++pinfo;
if (pinfo < this->buffer_end_)
++pinfo;
this->pinfo_ = pinfo;
return ret;
}
// class Dwarf_die
Dwarf_die::Dwarf_die(
Dwarf_info_reader* dwinfo,
off_t die_offset,
Dwarf_die* parent)
: dwinfo_(dwinfo), parent_(parent), die_offset_(die_offset),
child_offset_(0), sibling_offset_(0), abbrev_code_(NULL), attributes_(),
attributes_read_(false), name_(NULL), name_off_(-1), linkage_name_(NULL),
linkage_name_off_(-1), string_shndx_(0), specification_(0),
abstract_origin_(0)
{
size_t len;
const unsigned char* pdie = dwinfo->buffer_at_offset(die_offset);
if (pdie == NULL)
return;
unsigned int code = read_unsigned_LEB_128(pdie, &len);
if (code == 0)
{
if (parent != NULL)
parent->set_sibling_offset(die_offset + len);
return;
}
this->attr_offset_ = len;
// Lookup the abbrev code in the abbrev table.
this->abbrev_code_ = dwinfo->get_abbrev(code);
}
// Read all the attributes of the DIE.
bool
Dwarf_die::read_attributes()
{
if (this->attributes_read_)
return true;
gold_assert(this->abbrev_code_ != NULL);
const unsigned char* pdie =
this->dwinfo_->buffer_at_offset(this->die_offset_);
if (pdie == NULL)
return false;
const unsigned char* pattr = pdie + this->attr_offset_;
unsigned int nattr = this->abbrev_code_->attributes.size();
this->attributes_.reserve(nattr);
for (unsigned int i = 0; i < nattr; ++i)
{
size_t len;
unsigned int attr = this->abbrev_code_->attributes[i].attr;
unsigned int form = this->abbrev_code_->attributes[i].form;
if (form == elfcpp::DW_FORM_indirect)
{
form = read_unsigned_LEB_128(pattr, &len);
pattr += len;
}
off_t attr_off = this->die_offset_ + (pattr - pdie);
bool ref_form = false;
Attribute_value attr_value;
attr_value.attr = attr;
attr_value.form = form;
attr_value.aux.shndx = 0;
switch(form)
{
case elfcpp::DW_FORM_flag_present:
attr_value.val.intval = 1;
break;
case elfcpp::DW_FORM_strp:
{
off_t str_off;
if (this->dwinfo_->offset_size() == 4)
str_off = this->dwinfo_->read_from_pointer<32>(&pattr);
else
str_off = this->dwinfo_->read_from_pointer<64>(&pattr);
unsigned int shndx =
this->dwinfo_->lookup_reloc(attr_off, &str_off);
attr_value.aux.shndx = shndx;
attr_value.val.refval = str_off;
break;
}
case elfcpp::DW_FORM_sec_offset:
{
off_t sec_off;
if (this->dwinfo_->offset_size() == 4)
sec_off = this->dwinfo_->read_from_pointer<32>(&pattr);
else
sec_off = this->dwinfo_->read_from_pointer<64>(&pattr);
unsigned int shndx =
this->dwinfo_->lookup_reloc(attr_off, &sec_off);
attr_value.aux.shndx = shndx;
attr_value.val.refval = sec_off;
ref_form = true;
break;
}
case elfcpp::DW_FORM_addr:
case elfcpp::DW_FORM_ref_addr:
{
off_t sec_off;
if (this->dwinfo_->address_size() == 4)
sec_off = this->dwinfo_->read_from_pointer<32>(&pattr);
else
sec_off = this->dwinfo_->read_from_pointer<64>(&pattr);
unsigned int shndx =
this->dwinfo_->lookup_reloc(attr_off, &sec_off);
attr_value.aux.shndx = shndx;
attr_value.val.refval = sec_off;
ref_form = true;
break;
}
case elfcpp::DW_FORM_block1:
attr_value.aux.blocklen = *pattr++;
attr_value.val.blockval = pattr;
pattr += attr_value.aux.blocklen;
break;
case elfcpp::DW_FORM_block2:
attr_value.aux.blocklen =
this->dwinfo_->read_from_pointer<16>(&pattr);
attr_value.val.blockval = pattr;
pattr += attr_value.aux.blocklen;
break;
case elfcpp::DW_FORM_block4:
attr_value.aux.blocklen =
this->dwinfo_->read_from_pointer<32>(&pattr);
attr_value.val.blockval = pattr;
pattr += attr_value.aux.blocklen;
break;
case elfcpp::DW_FORM_block:
case elfcpp::DW_FORM_exprloc:
attr_value.aux.blocklen = read_unsigned_LEB_128(pattr, &len);
attr_value.val.blockval = pattr + len;
pattr += len + attr_value.aux.blocklen;
break;
case elfcpp::DW_FORM_data1:
case elfcpp::DW_FORM_flag:
attr_value.val.intval = *pattr++;
break;
case elfcpp::DW_FORM_ref1:
attr_value.val.refval = *pattr++;
ref_form = true;
break;
case elfcpp::DW_FORM_data2:
attr_value.val.intval =
this->dwinfo_->read_from_pointer<16>(&pattr);
break;
case elfcpp::DW_FORM_ref2:
attr_value.val.refval =
this->dwinfo_->read_from_pointer<16>(&pattr);
ref_form = true;
break;
case elfcpp::DW_FORM_data4:
{
off_t sec_off;
sec_off = this->dwinfo_->read_from_pointer<32>(&pattr);
unsigned int shndx =
this->dwinfo_->lookup_reloc(attr_off, &sec_off);
attr_value.aux.shndx = shndx;
attr_value.val.intval = sec_off;
break;
}
case elfcpp::DW_FORM_ref4:
{
off_t sec_off;
sec_off = this->dwinfo_->read_from_pointer<32>(&pattr);
unsigned int shndx =
this->dwinfo_->lookup_reloc(attr_off, &sec_off);
attr_value.aux.shndx = shndx;
attr_value.val.refval = sec_off;
ref_form = true;
break;
}
case elfcpp::DW_FORM_data8:
{
off_t sec_off;
sec_off = this->dwinfo_->read_from_pointer<64>(&pattr);
unsigned int shndx =
this->dwinfo_->lookup_reloc(attr_off, &sec_off);
attr_value.aux.shndx = shndx;
attr_value.val.intval = sec_off;
break;
}
case elfcpp::DW_FORM_ref_sig8:
attr_value.val.uintval =
this->dwinfo_->read_from_pointer<64>(&pattr);
break;
case elfcpp::DW_FORM_ref8:
{
off_t sec_off;
sec_off = this->dwinfo_->read_from_pointer<64>(&pattr);
unsigned int shndx =
this->dwinfo_->lookup_reloc(attr_off, &sec_off);
attr_value.aux.shndx = shndx;
attr_value.val.refval = sec_off;
ref_form = true;
break;
}
case elfcpp::DW_FORM_ref_udata:
attr_value.val.refval = read_unsigned_LEB_128(pattr, &len);
ref_form = true;
pattr += len;
break;
case elfcpp::DW_FORM_udata:
case elfcpp::DW_FORM_GNU_addr_index:
case elfcpp::DW_FORM_GNU_str_index:
attr_value.val.uintval = read_unsigned_LEB_128(pattr, &len);
pattr += len;
break;
case elfcpp::DW_FORM_sdata:
attr_value.val.intval = read_signed_LEB_128(pattr, &len);
pattr += len;
break;
case elfcpp::DW_FORM_string:
attr_value.val.stringval = reinterpret_cast<const char*>(pattr);
len = strlen(attr_value.val.stringval);
pattr += len + 1;
break;
default:
return false;
}
// Cache the most frequently-requested attributes.
switch (attr)
{
case elfcpp::DW_AT_name:
if (form == elfcpp::DW_FORM_string)
this->name_ = attr_value.val.stringval;
else if (form == elfcpp::DW_FORM_strp)
{
// All indirect strings should refer to the same
// string section, so we just save the last one seen.
this->string_shndx_ = attr_value.aux.shndx;
this->name_off_ = attr_value.val.refval;
}
break;
case elfcpp::DW_AT_linkage_name:
case elfcpp::DW_AT_MIPS_linkage_name:
if (form == elfcpp::DW_FORM_string)
this->linkage_name_ = attr_value.val.stringval;
else if (form == elfcpp::DW_FORM_strp)
{
// All indirect strings should refer to the same
// string section, so we just save the last one seen.
this->string_shndx_ = attr_value.aux.shndx;
this->linkage_name_off_ = attr_value.val.refval;
}
break;
case elfcpp::DW_AT_specification:
if (ref_form)
this->specification_ = attr_value.val.refval;
break;
case elfcpp::DW_AT_abstract_origin:
if (ref_form)
this->abstract_origin_ = attr_value.val.refval;
break;
case elfcpp::DW_AT_sibling:
if (ref_form && attr_value.aux.shndx == 0)
this->sibling_offset_ = attr_value.val.refval;
default:
break;
}
this->attributes_.push_back(attr_value);
}
// Now that we know where the next DIE begins, record the offset
// to avoid later recalculation.
if (this->has_children())
this->child_offset_ = this->die_offset_ + (pattr - pdie);
else
this->sibling_offset_ = this->die_offset_ + (pattr - pdie);
this->attributes_read_ = true;
return true;
}
// Skip all the attributes of the DIE and return the offset of the next DIE.
off_t
Dwarf_die::skip_attributes()
{
gold_assert(this->abbrev_code_ != NULL);
const unsigned char* pdie =
this->dwinfo_->buffer_at_offset(this->die_offset_);
if (pdie == NULL)
return 0;
const unsigned char* pattr = pdie + this->attr_offset_;
for (unsigned int i = 0; i < this->abbrev_code_->attributes.size(); ++i)
{
size_t len;
unsigned int form = this->abbrev_code_->attributes[i].form;
if (form == elfcpp::DW_FORM_indirect)
{
form = read_unsigned_LEB_128(pattr, &len);
pattr += len;
}
switch(form)
{
case elfcpp::DW_FORM_flag_present:
break;
case elfcpp::DW_FORM_strp:
case elfcpp::DW_FORM_sec_offset:
pattr += this->dwinfo_->offset_size();
break;
case elfcpp::DW_FORM_addr:
case elfcpp::DW_FORM_ref_addr:
pattr += this->dwinfo_->address_size();
break;
case elfcpp::DW_FORM_block1:
pattr += 1 + *pattr;
break;
case elfcpp::DW_FORM_block2:
{
uint16_t block_size;
block_size = this->dwinfo_->read_from_pointer<16>(&pattr);
pattr += block_size;
break;
}
case elfcpp::DW_FORM_block4:
{
uint32_t block_size;
block_size = this->dwinfo_->read_from_pointer<32>(&pattr);
pattr += block_size;
break;
}
case elfcpp::DW_FORM_block:
case elfcpp::DW_FORM_exprloc:
{
uint64_t block_size;
block_size = read_unsigned_LEB_128(pattr, &len);
pattr += len + block_size;
break;
}
case elfcpp::DW_FORM_data1:
case elfcpp::DW_FORM_ref1:
case elfcpp::DW_FORM_flag:
pattr += 1;
break;
case elfcpp::DW_FORM_data2:
case elfcpp::DW_FORM_ref2:
pattr += 2;
break;
case elfcpp::DW_FORM_data4:
case elfcpp::DW_FORM_ref4:
pattr += 4;
break;
case elfcpp::DW_FORM_data8:
case elfcpp::DW_FORM_ref8:
case elfcpp::DW_FORM_ref_sig8:
pattr += 8;
break;
case elfcpp::DW_FORM_ref_udata:
case elfcpp::DW_FORM_udata:
case elfcpp::DW_FORM_GNU_addr_index:
case elfcpp::DW_FORM_GNU_str_index:
read_unsigned_LEB_128(pattr, &len);
pattr += len;
break;
case elfcpp::DW_FORM_sdata:
read_signed_LEB_128(pattr, &len);
pattr += len;
break;
case elfcpp::DW_FORM_string:
len = strlen(reinterpret_cast<const char*>(pattr));
pattr += len + 1;
break;
default:
return 0;
}
}
return this->die_offset_ + (pattr - pdie);
}
// Get the name of the DIE and cache it.
void
Dwarf_die::set_name()
{
if (this->name_ != NULL || !this->read_attributes())
return;
if (this->name_off_ != -1)
this->name_ = this->dwinfo_->get_string(this->name_off_,
this->string_shndx_);
}
// Get the linkage name of the DIE and cache it.
void
Dwarf_die::set_linkage_name()
{
if (this->linkage_name_ != NULL || !this->read_attributes())
return;
if (this->linkage_name_off_ != -1)
this->linkage_name_ = this->dwinfo_->get_string(this->linkage_name_off_,
this->string_shndx_);
}
// Return the value of attribute ATTR.
const Dwarf_die::Attribute_value*
Dwarf_die::attribute(unsigned int attr)
{
if (!this->read_attributes())
return NULL;
for (unsigned int i = 0; i < this->attributes_.size(); ++i)
{
if (this->attributes_[i].attr == attr)
return &this->attributes_[i];
}
return NULL;
}
const char*
Dwarf_die::string_attribute(unsigned int attr)
{
const Attribute_value* attr_val = this->attribute(attr);
if (attr_val == NULL)
return NULL;
switch (attr_val->form)
{
case elfcpp::DW_FORM_string:
return attr_val->val.stringval;
case elfcpp::DW_FORM_strp:
return this->dwinfo_->get_string(attr_val->val.refval,
attr_val->aux.shndx);
default:
return NULL;
}
}
int64_t
Dwarf_die::int_attribute(unsigned int attr)
{
const Attribute_value* attr_val = this->attribute(attr);
if (attr_val == NULL)
return 0;
switch (attr_val->form)
{
case elfcpp::DW_FORM_flag_present:
case elfcpp::DW_FORM_data1:
case elfcpp::DW_FORM_flag:
case elfcpp::DW_FORM_data2:
case elfcpp::DW_FORM_data4:
case elfcpp::DW_FORM_data8:
case elfcpp::DW_FORM_sdata:
return attr_val->val.intval;
default:
return 0;
}
}
uint64_t
Dwarf_die::uint_attribute(unsigned int attr)
{
const Attribute_value* attr_val = this->attribute(attr);
if (attr_val == NULL)
return 0;
switch (attr_val->form)
{
case elfcpp::DW_FORM_flag_present:
case elfcpp::DW_FORM_data1:
case elfcpp::DW_FORM_flag:
case elfcpp::DW_FORM_data4:
case elfcpp::DW_FORM_data8:
case elfcpp::DW_FORM_ref_sig8:
case elfcpp::DW_FORM_udata:
return attr_val->val.uintval;
default:
return 0;
}
}
off_t
Dwarf_die::ref_attribute(unsigned int attr, unsigned int* shndx)
{
const Attribute_value* attr_val = this->attribute(attr);
if (attr_val == NULL)
return -1;
switch (attr_val->form)
{
case elfcpp::DW_FORM_sec_offset:
case elfcpp::DW_FORM_addr:
case elfcpp::DW_FORM_ref_addr:
case elfcpp::DW_FORM_ref1:
case elfcpp::DW_FORM_ref2:
case elfcpp::DW_FORM_ref4:
case elfcpp::DW_FORM_ref8:
case elfcpp::DW_FORM_ref_udata:
*shndx = attr_val->aux.shndx;
return attr_val->val.refval;
case elfcpp::DW_FORM_ref_sig8:
*shndx = attr_val->aux.shndx;
return attr_val->val.uintval;
case elfcpp::DW_FORM_data4:
case elfcpp::DW_FORM_data8:
*shndx = attr_val->aux.shndx;
return attr_val->val.intval;
default:
return -1;
}
}
off_t
Dwarf_die::address_attribute(unsigned int attr, unsigned int* shndx)
{
const Attribute_value* attr_val = this->attribute(attr);
if (attr_val == NULL || attr_val->form != elfcpp::DW_FORM_addr)
return -1;
*shndx = attr_val->aux.shndx;
return attr_val->val.refval;
}
// Return the offset of this DIE's first child.
off_t
Dwarf_die::child_offset()
{
gold_assert(this->abbrev_code_ != NULL);
if (!this->has_children())
return 0;
if (this->child_offset_ == 0)
this->child_offset_ = this->skip_attributes();
return this->child_offset_;
}
// Return the offset of this DIE's next sibling.
off_t
Dwarf_die::sibling_offset()
{
gold_assert(this->abbrev_code_ != NULL);
if (this->sibling_offset_ != 0)
return this->sibling_offset_;
if (!this->has_children())
{
this->sibling_offset_ = this->skip_attributes();
return this->sibling_offset_;
}
if (this->has_sibling_attribute())
{
if (!this->read_attributes())
return 0;
if (this->sibling_offset_ != 0)
return this->sibling_offset_;
}
// Skip over the children.
off_t child_offset = this->child_offset();
while (child_offset > 0)
{
Dwarf_die die(this->dwinfo_, child_offset, this);
// The Dwarf_die ctor will set this DIE's sibling offset
// when it reads a zero abbrev code.
if (die.tag() == 0)
break;
child_offset = die.sibling_offset();
}
// This should be set by now. If not, there was a problem reading
// the DWARF info, and we return 0.
return this->sibling_offset_;
}
// class Dwarf_info_reader
// Check that the pointer P is within the current compilation unit.
inline bool
Dwarf_info_reader::check_buffer(const unsigned char* p) const
{
if (p > this->buffer_ + this->cu_offset_ + this->cu_length_)
{
gold_warning(_("%s: corrupt debug info in %s"),
this->object_->name().c_str(),
this->object_->section_name(this->shndx_).c_str());
return false;
}
return true;
}
// Begin parsing the debug info. This calls visit_compilation_unit()
// or visit_type_unit() for each compilation or type unit found in the
// section, and visit_die() for each top-level DIE.
void
Dwarf_info_reader::parse()
{
if (this->object_->is_big_endian())
{
#if defined(HAVE_TARGET_32_BIG) || defined(HAVE_TARGET_64_BIG)
this->do_parse<true>();
#else
gold_unreachable();
#endif
}
else
{
#if defined(HAVE_TARGET_32_LITTLE) || defined(HAVE_TARGET_64_LITTLE)
this->do_parse<false>();
#else
gold_unreachable();
#endif
}
}
template<bool big_endian>
void
Dwarf_info_reader::do_parse()
{
// Get the section contents and decompress if necessary.
section_size_type buffer_size;
bool buffer_is_new;
this->buffer_ = this->object_->decompressed_section_contents(this->shndx_,
&buffer_size,
&buffer_is_new);
if (this->buffer_ == NULL || buffer_size == 0)
return;
this->buffer_end_ = this->buffer_ + buffer_size;
// The offset of this input section in the output section.
off_t section_offset = this->object_->output_section_offset(this->shndx_);
// Start tracking relocations for this section.
this->reloc_mapper_ = make_elf_reloc_mapper(this->object_, this->symtab_,
this->symtab_size_);
this->reloc_mapper_->initialize(this->reloc_shndx_, this->reloc_type_);
// Loop over compilation units (or type units).
unsigned int abbrev_shndx = this->abbrev_shndx_;
off_t abbrev_offset = 0;
const unsigned char* pinfo = this->buffer_;
while (pinfo < this->buffer_end_)
{
// Read the compilation (or type) unit header.
const unsigned char* cu_start = pinfo;
this->cu_offset_ = cu_start - this->buffer_;
this->cu_length_ = this->buffer_end_ - cu_start;
// Read unit_length (4 or 12 bytes).
if (!this->check_buffer(pinfo + 4))
break;
uint32_t unit_length =
elfcpp::Swap_unaligned<32, big_endian>::readval(pinfo);
pinfo += 4;
if (unit_length == 0xffffffff)
{
if (!this->check_buffer(pinfo + 8))
break;
unit_length = elfcpp::Swap_unaligned<64, big_endian>::readval(pinfo);
pinfo += 8;
this->offset_size_ = 8;
}
else
this->offset_size_ = 4;
if (!this->check_buffer(pinfo + unit_length))
break;
const unsigned char* cu_end = pinfo + unit_length;
this->cu_length_ = cu_end - cu_start;
if (!this->check_buffer(pinfo + 2 + this->offset_size_ + 1))
break;
// Read version (2 bytes).
this->cu_version_ =
elfcpp::Swap_unaligned<16, big_endian>::readval(pinfo);
pinfo += 2;
// Read debug_abbrev_offset (4 or 8 bytes).
if (this->offset_size_ == 4)
abbrev_offset = elfcpp::Swap_unaligned<32, big_endian>::readval(pinfo);
else
abbrev_offset = elfcpp::Swap_unaligned<64, big_endian>::readval(pinfo);
if (this->reloc_shndx_ > 0)
{
off_t reloc_offset = pinfo - this->buffer_;
off_t value;
abbrev_shndx =
this->reloc_mapper_->get_reloc_target(reloc_offset, &value);
if (abbrev_shndx == 0)
return;
if (this->reloc_type_ == elfcpp::SHT_REL)
abbrev_offset += value;
else
abbrev_offset = value;
}
pinfo += this->offset_size_;
// Read address_size (1 byte).
this->address_size_ = *pinfo++;
// For type units, read the two extra fields.
uint64_t signature = 0;
off_t type_offset = 0;
if (this->is_type_unit_)
{
if (!this->check_buffer(pinfo + 8 + this->offset_size_))
break;
// Read type_signature (8 bytes).
signature = elfcpp::Swap_unaligned<64, big_endian>::readval(pinfo);
pinfo += 8;
// Read type_offset (4 or 8 bytes).
if (this->offset_size_ == 4)
type_offset =
elfcpp::Swap_unaligned<32, big_endian>::readval(pinfo);
else
type_offset =
elfcpp::Swap_unaligned<64, big_endian>::readval(pinfo);
pinfo += this->offset_size_;
}
// Read the .debug_abbrev table.
this->abbrev_table_.read_abbrevs(this->object_, abbrev_shndx,
abbrev_offset);
// Visit the root DIE.
Dwarf_die root_die(this,
pinfo - (this->buffer_ + this->cu_offset_),
NULL);
if (root_die.tag() != 0)
{
// Visit the CU or TU.
if (this->is_type_unit_)
this->visit_type_unit(section_offset + this->cu_offset_,
type_offset, signature, &root_die);
else
this->visit_compilation_unit(section_offset + this->cu_offset_,
cu_end - cu_start, &root_die);
}
// Advance to the next CU.
pinfo = cu_end;
}
if (buffer_is_new)
{
delete[] this->buffer_;
this->buffer_ = NULL;
}
}
// Read the DWARF string table.
bool
Dwarf_info_reader::do_read_string_table(unsigned int string_shndx)
{
Relobj* object = this->object_;
// If we don't have relocations, string_shndx will be 0, and
// we'll have to hunt for the .debug_str section.
if (string_shndx == 0)
{
for (unsigned int i = 1; i < this->object_->shnum(); ++i)
{
std::string name = object->section_name(i);
if (name == ".debug_str")
{
string_shndx = i;
this->string_output_section_offset_ =
object->output_section_offset(i);
break;
}
}
if (string_shndx == 0)
return false;
}
if (this->owns_string_buffer_ && this->string_buffer_ != NULL)
{
delete[] this->string_buffer_;
this->owns_string_buffer_ = false;
}
// Get the secton contents and decompress if necessary.
section_size_type buffer_size;
const unsigned char* buffer =
object->decompressed_section_contents(string_shndx,
&buffer_size,
&this->owns_string_buffer_);
this->string_buffer_ = reinterpret_cast<const char*>(buffer);
this->string_buffer_end_ = this->string_buffer_ + buffer_size;
this->string_shndx_ = string_shndx;
return true;
}
// Read a possibly unaligned integer of SIZE.
template <int valsize>
inline typename elfcpp::Valtype_base<valsize>::Valtype
Dwarf_info_reader::read_from_pointer(const unsigned char* source)
{
typename elfcpp::Valtype_base<valsize>::Valtype return_value;
if (this->object_->is_big_endian())
return_value = elfcpp::Swap_unaligned<valsize, true>::readval(source);
else
return_value = elfcpp::Swap_unaligned<valsize, false>::readval(source);
return return_value;
}
// Read a possibly unaligned integer of SIZE. Update SOURCE after read.
template <int valsize>
inline typename elfcpp::Valtype_base<valsize>::Valtype
Dwarf_info_reader::read_from_pointer(const unsigned char** source)
{
typename elfcpp::Valtype_base<valsize>::Valtype return_value;
if (this->object_->is_big_endian())
return_value = elfcpp::Swap_unaligned<valsize, true>::readval(*source);
else
return_value = elfcpp::Swap_unaligned<valsize, false>::readval(*source);
*source += valsize / 8;
return return_value;
}
// Look for a relocation at offset ATTR_OFF in the dwarf info,
// and return the section index and offset of the target.
unsigned int
Dwarf_info_reader::lookup_reloc(off_t attr_off, off_t* target_off)
{
off_t value;
attr_off += this->cu_offset_;
unsigned int shndx = this->reloc_mapper_->get_reloc_target(attr_off, &value);
if (shndx == 0)
return 0;
if (this->reloc_type_ == elfcpp::SHT_REL)
*target_off += value;
else
*target_off = value;
return shndx;
}
// Return a string from the DWARF string table.
const char*
Dwarf_info_reader::get_string(off_t str_off, unsigned int string_shndx)
{
if (!this->read_string_table(string_shndx))
return NULL;
// Correct the offset. For incremental update links, we have a
// relocated offset that is relative to the output section, but
// here we need an offset relative to the input section.
str_off -= this->string_output_section_offset_;
const char* p = this->string_buffer_ + str_off;
if (p < this->string_buffer_ || p >= this->string_buffer_end_)
return NULL;
return p;
}
// The following are default, do-nothing, implementations of the
// hook methods normally provided by a derived class. We provide
// default implementations rather than no implementation so that
// a derived class needs to implement only the hooks that it needs
// to use.
// Process a compilation unit and parse its child DIE.
void
Dwarf_info_reader::visit_compilation_unit(off_t, off_t, Dwarf_die*)
{
}
// Process a type unit and parse its child DIE.
void
Dwarf_info_reader::visit_type_unit(off_t, off_t, uint64_t, Dwarf_die*)
{
}
// class Sized_dwarf_line_info
struct LineStateMachine
{
int file_num;
uint64_t address;
int line_num;
int column_num;
unsigned int shndx; // the section address refers to
bool is_stmt; // stmt means statement.
bool basic_block;
bool end_sequence;
};
static void
ResetLineStateMachine(struct LineStateMachine* lsm, bool default_is_stmt)
{
lsm->file_num = 1;
lsm->address = 0;
lsm->line_num = 1;
lsm->column_num = 0;
lsm->shndx = -1U;
lsm->is_stmt = default_is_stmt;
lsm->basic_block = false;
lsm->end_sequence = false;
}
template<int size, bool big_endian>
Sized_dwarf_line_info<size, big_endian>::Sized_dwarf_line_info(
Object* object,
unsigned int read_shndx)
: data_valid_(false), buffer_(NULL), buffer_start_(NULL),
reloc_mapper_(NULL), symtab_buffer_(NULL), directories_(), files_(),
current_header_index_(-1)
{
unsigned int debug_shndx;
for (debug_shndx = 1; debug_shndx < object->shnum(); ++debug_shndx)
{
// FIXME: do this more efficiently: section_name() isn't super-fast
std::string name = object->section_name(debug_shndx);
if (name == ".debug_line" || name == ".zdebug_line")
{
section_size_type buffer_size;
bool is_new = false;
this->buffer_ = object->decompressed_section_contents(debug_shndx,
&buffer_size,
&is_new);
if (is_new)
this->buffer_start_ = this->buffer_;
this->buffer_end_ = this->buffer_ + buffer_size;
break;
}
}
if (this->buffer_ == NULL)
return;
// Find the relocation section for ".debug_line".
// We expect these for relobjs (.o's) but not dynobjs (.so's).
unsigned int reloc_shndx = 0;
for (unsigned int i = 0; i < object->shnum(); ++i)
{
unsigned int reloc_sh_type = object->section_type(i);
if ((reloc_sh_type == elfcpp::SHT_REL
|| reloc_sh_type == elfcpp::SHT_RELA)
&& object->section_info(i) == debug_shndx)
{
reloc_shndx = i;
this->track_relocs_type_ = reloc_sh_type;
break;
}
}
// Finally, we need the symtab section to interpret the relocs.
if (reloc_shndx != 0)
{
unsigned int symtab_shndx;
for (symtab_shndx = 0; symtab_shndx < object->shnum(); ++symtab_shndx)
if (object->section_type(symtab_shndx) == elfcpp::SHT_SYMTAB)
{
this->symtab_buffer_ = object->section_contents(
symtab_shndx, &this->symtab_buffer_size_, false);
break;
}
if (this->symtab_buffer_ == NULL)
return;
}
this->reloc_mapper_ =
new Sized_elf_reloc_mapper<size, big_endian>(object,
this->symtab_buffer_,
this->symtab_buffer_size_);
if (!this->reloc_mapper_->initialize(reloc_shndx, this->track_relocs_type_))
return;
// Now that we have successfully read all the data, parse the debug
// info.
this->data_valid_ = true;
this->read_line_mappings(read_shndx);
}
// Read the DWARF header.
template<int size, bool big_endian>
const unsigned char*
Sized_dwarf_line_info<size, big_endian>::read_header_prolog(
const unsigned char* lineptr)
{
uint32_t initial_length = elfcpp::Swap_unaligned<32, big_endian>::readval(lineptr);
lineptr += 4;
// In DWARF2/3, if the initial length is all 1 bits, then the offset
// size is 8 and we need to read the next 8 bytes for the real length.
if (initial_length == 0xffffffff)
{
header_.offset_size = 8;
initial_length = elfcpp::Swap_unaligned<64, big_endian>::readval(lineptr);
lineptr += 8;
}
else
header_.offset_size = 4;
header_.total_length = initial_length;
gold_assert(lineptr + header_.total_length <= buffer_end_);
header_.version = elfcpp::Swap_unaligned<16, big_endian>::readval(lineptr);
lineptr += 2;
if (header_.offset_size == 4)
header_.prologue_length = elfcpp::Swap_unaligned<32, big_endian>::readval(lineptr);
else
header_.prologue_length = elfcpp::Swap_unaligned<64, big_endian>::readval(lineptr);
lineptr += header_.offset_size;
header_.min_insn_length = *lineptr;
lineptr += 1;
header_.default_is_stmt = *lineptr;
lineptr += 1;
header_.line_base = *reinterpret_cast<const signed char*>(lineptr);
lineptr += 1;
header_.line_range = *lineptr;
lineptr += 1;
header_.opcode_base = *lineptr;
lineptr += 1;
header_.std_opcode_lengths.resize(header_.opcode_base + 1);
header_.std_opcode_lengths[0] = 0;
for (int i = 1; i < header_.opcode_base; i++)
{
header_.std_opcode_lengths[i] = *lineptr;
lineptr += 1;
}
return lineptr;
}
// The header for a debug_line section is mildly complicated, because
// the line info is very tightly encoded.
template<int size, bool big_endian>
const unsigned char*
Sized_dwarf_line_info<size, big_endian>::read_header_tables(
const unsigned char* lineptr)
{
++this->current_header_index_;
// Create a new directories_ entry and a new files_ entry for our new
// header. We initialize each with a single empty element, because
// dwarf indexes directory and filenames starting at 1.
gold_assert(static_cast<int>(this->directories_.size())
== this->current_header_index_);
gold_assert(static_cast<int>(this->files_.size())
== this->current_header_index_);
this->directories_.push_back(std::vector<std::string>(1));
this->files_.push_back(std::vector<std::pair<int, std::string> >(1));
// It is legal for the directory entry table to be empty.
if (*lineptr)
{
int dirindex = 1;
while (*lineptr)
{
const char* dirname = reinterpret_cast<const char*>(lineptr);
gold_assert(dirindex
== static_cast<int>(this->directories_.back().size()));
this->directories_.back().push_back(dirname);
lineptr += this->directories_.back().back().size() + 1;
dirindex++;
}
}
lineptr++;
// It is also legal for the file entry table to be empty.
if (*lineptr)
{
int fileindex = 1;
size_t len;
while (*lineptr)
{
const char* filename = reinterpret_cast<const char*>(lineptr);
lineptr += strlen(filename) + 1;
uint64_t dirindex = read_unsigned_LEB_128(lineptr, &len);
lineptr += len;
if (dirindex >= this->directories_.back().size())
dirindex = 0;
int dirindexi = static_cast<int>(dirindex);
read_unsigned_LEB_128(lineptr, &len); // mod_time
lineptr += len;
read_unsigned_LEB_128(lineptr, &len); // filelength
lineptr += len;
gold_assert(fileindex
== static_cast<int>(this->files_.back().size()));
this->files_.back().push_back(std::make_pair(dirindexi, filename));
fileindex++;
}
}
lineptr++;
return lineptr;
}
// Process a single opcode in the .debug.line structure.
template<int size, bool big_endian>
bool
Sized_dwarf_line_info<size, big_endian>::process_one_opcode(
const unsigned char* start, struct LineStateMachine* lsm, size_t* len)
{
size_t oplen = 0;
size_t templen;
unsigned char opcode = *start;
oplen++;
start++;
// If the opcode is great than the opcode_base, it is a special
// opcode. Most line programs consist mainly of special opcodes.
if (opcode >= header_.opcode_base)
{
opcode -= header_.opcode_base;
const int advance_address = ((opcode / header_.line_range)
* header_.min_insn_length);
lsm->address += advance_address;
const int advance_line = ((opcode % header_.line_range)
+ header_.line_base);
lsm->line_num += advance_line;
lsm->basic_block = true;
*len = oplen;
return true;
}
// Otherwise, we have the regular opcodes
switch (opcode)
{
case elfcpp::DW_LNS_copy:
lsm->basic_block = false;
*len = oplen;
return true;
case elfcpp::DW_LNS_advance_pc:
{
const uint64_t advance_address
= read_unsigned_LEB_128(start, &templen);
oplen += templen;
lsm->address += header_.min_insn_length * advance_address;
}
break;
case elfcpp::DW_LNS_advance_line:
{
const uint64_t advance_line = read_signed_LEB_128(start, &templen);
oplen += templen;
lsm->line_num += advance_line;
}
break;
case elfcpp::DW_LNS_set_file:
{
const uint64_t fileno = read_unsigned_LEB_128(start, &templen);
oplen += templen;
lsm->file_num = fileno;
}
break;
case elfcpp::DW_LNS_set_column:
{
const uint64_t colno = read_unsigned_LEB_128(start, &templen);
oplen += templen;
lsm->column_num = colno;
}
break;
case elfcpp::DW_LNS_negate_stmt:
lsm->is_stmt = !lsm->is_stmt;
break;
case elfcpp::DW_LNS_set_basic_block:
lsm->basic_block = true;
break;
case elfcpp::DW_LNS_fixed_advance_pc:
{
int advance_address;
advance_address = elfcpp::Swap_unaligned<16, big_endian>::readval(start);
oplen += 2;
lsm->address += advance_address;
}
break;
case elfcpp::DW_LNS_const_add_pc:
{
const int advance_address = (header_.min_insn_length
* ((255 - header_.opcode_base)
/ header_.line_range));
lsm->address += advance_address;
}
break;
case elfcpp::DW_LNS_extended_op:
{
const uint64_t extended_op_len
= read_unsigned_LEB_128(start, &templen);
start += templen;
oplen += templen + extended_op_len;
const unsigned char extended_op = *start;
start++;
switch (extended_op)
{
case elfcpp::DW_LNE_end_sequence:
// This means that the current byte is the one immediately
// after a set of instructions. Record the current line
// for up to one less than the current address.
lsm->line_num = -1;
lsm->end_sequence = true;
*len = oplen;
return true;
case elfcpp::DW_LNE_set_address:
{
lsm->address =
elfcpp::Swap_unaligned<size, big_endian>::readval(start);
typename Reloc_map::const_iterator it
= this->reloc_map_.find(start - this->buffer_);
if (it != reloc_map_.end())
{
// If this is a SHT_RELA section, then ignore the
// section contents. This assumes that this is a
// straight reloc which just uses the reloc addend.
// The reloc addend has already been included in the
// symbol value.
if (this->track_relocs_type_ == elfcpp::SHT_RELA)
lsm->address = 0;
// Add in the symbol value.
lsm->address += it->second.second;
lsm->shndx = it->second.first;
}
else
{
// If we're a normal .o file, with relocs, every
// set_address should have an associated relocation.
if (this->input_is_relobj())
this->data_valid_ = false;
}
break;
}
case elfcpp::DW_LNE_define_file:
{
const char* filename = reinterpret_cast<const char*>(start);
templen = strlen(filename) + 1;
start += templen;
uint64_t dirindex = read_unsigned_LEB_128(start, &templen);
if (dirindex >= this->directories_.back().size())
dirindex = 0;
int dirindexi = static_cast<int>(dirindex);
// This opcode takes two additional ULEB128 parameters
// (mod_time and filelength), but we don't use those
// values. Because OPLEN already tells us how far to
// skip to the next opcode, we don't need to read
// them at all.
this->files_.back().push_back(std::make_pair(dirindexi,
filename));
}
break;
}
}
break;
default:
{
// Ignore unknown opcode silently
for (int i = 0; i < header_.std_opcode_lengths[opcode]; i++)
{
size_t templen;
read_unsigned_LEB_128(start, &templen);
start += templen;
oplen += templen;
}
}
break;
}
*len = oplen;
return false;
}
// Read the debug information at LINEPTR and store it in the line
// number map.
template<int size, bool big_endian>
unsigned const char*
Sized_dwarf_line_info<size, big_endian>::read_lines(unsigned const char* lineptr,
unsigned int shndx)
{
struct LineStateMachine lsm;
// LENGTHSTART is the place the length field is based on. It is the
// point in the header after the initial length field.
const unsigned char* lengthstart = buffer_;
// In 64 bit dwarf, the initial length is 12 bytes, because of the
// 0xffffffff at the start.
if (header_.offset_size == 8)
lengthstart += 12;
else
lengthstart += 4;
while (lineptr < lengthstart + header_.total_length)
{
ResetLineStateMachine(&lsm, header_.default_is_stmt);
while (!lsm.end_sequence)
{
size_t oplength;
bool add_line = this->process_one_opcode(lineptr, &lsm, &oplength);
if (add_line
&& (shndx == -1U || lsm.shndx == -1U || shndx == lsm.shndx))
{
Offset_to_lineno_entry entry
= { static_cast<off_t>(lsm.address),
this->current_header_index_,
static_cast<unsigned int>(lsm.file_num),
true, lsm.line_num };
std::vector<Offset_to_lineno_entry>&
map(this->line_number_map_[lsm.shndx]);
// If we see two consecutive entries with the same
// offset and a real line number, then mark the first
// one as non-canonical.
if (!map.empty()
&& (map.back().offset == static_cast<off_t>(lsm.address))
&& lsm.line_num != -1
&& map.back().line_num != -1)
map.back().last_line_for_offset = false;
map.push_back(entry);
}
lineptr += oplength;
}
}
return lengthstart + header_.total_length;
}
// Read the relocations into a Reloc_map.
template<int size, bool big_endian>
void
Sized_dwarf_line_info<size, big_endian>::read_relocs()
{
if (this->symtab_buffer_ == NULL)
return;
off_t value;
off_t reloc_offset;
while ((reloc_offset = this->reloc_mapper_->next_offset()) != -1)
{
const unsigned int shndx =
this->reloc_mapper_->get_reloc_target(reloc_offset, &value);
// There is no reason to record non-ordinary section indexes, or
// SHN_UNDEF, because they will never match the real section.
if (shndx != 0)
this->reloc_map_[reloc_offset] = std::make_pair(shndx, value);
this->reloc_mapper_->advance(reloc_offset + 1);
}
}
// Read the line number info.
template<int size, bool big_endian>
void
Sized_dwarf_line_info<size, big_endian>::read_line_mappings(unsigned int shndx)
{
gold_assert(this->data_valid_ == true);
this->read_relocs();
while (this->buffer_ < this->buffer_end_)
{
const unsigned char* lineptr = this->buffer_;
lineptr = this->read_header_prolog(lineptr);
lineptr = this->read_header_tables(lineptr);
lineptr = this->read_lines(lineptr, shndx);
this->buffer_ = lineptr;
}
// Sort the lines numbers, so addr2line can use binary search.
for (typename Lineno_map::iterator it = line_number_map_.begin();
it != line_number_map_.end();
++it)
// Each vector needs to be sorted by offset.
std::sort(it->second.begin(), it->second.end());
}
// Some processing depends on whether the input is a .o file or not.
// For instance, .o files have relocs, and have .debug_lines
// information on a per section basis. .so files, on the other hand,
// lack relocs, and offsets are unique, so we can ignore the section
// information.
template<int size, bool big_endian>
bool
Sized_dwarf_line_info<size, big_endian>::input_is_relobj()
{
// Only .o files have relocs and the symtab buffer that goes with them.
return this->symtab_buffer_ != NULL;
}
// Given an Offset_to_lineno_entry vector, and an offset, figure out
// if the offset points into a function according to the vector (see
// comments below for the algorithm). If it does, return an iterator
// into the vector that points to the line-number that contains that
// offset. If not, it returns vector::end().
static std::vector<Offset_to_lineno_entry>::const_iterator
offset_to_iterator(const std::vector<Offset_to_lineno_entry>* offsets,
off_t offset)
{
const Offset_to_lineno_entry lookup_key = { offset, 0, 0, true, 0 };
// lower_bound() returns the smallest offset which is >= lookup_key.
// If no offset in offsets is >= lookup_key, returns end().
std::vector<Offset_to_lineno_entry>::const_iterator it
= std::lower_bound(offsets->begin(), offsets->end(), lookup_key);
// This code is easiest to understand with a concrete example.
// Here's a possible offsets array:
// {{offset = 3211, header_num = 0, file_num = 1, last, line_num = 16}, // 0
// {offset = 3224, header_num = 0, file_num = 1, last, line_num = 20}, // 1
// {offset = 3226, header_num = 0, file_num = 1, last, line_num = 22}, // 2
// {offset = 3231, header_num = 0, file_num = 1, last, line_num = 25}, // 3
// {offset = 3232, header_num = 0, file_num = 1, last, line_num = -1}, // 4
// {offset = 3232, header_num = 0, file_num = 1, last, line_num = 65}, // 5
// {offset = 3235, header_num = 0, file_num = 1, last, line_num = 66}, // 6
// {offset = 3236, header_num = 0, file_num = 1, last, line_num = -1}, // 7
// {offset = 5764, header_num = 0, file_num = 1, last, line_num = 48}, // 8
// {offset = 5764, header_num = 0, file_num = 1,!last, line_num = 47}, // 9
// {offset = 5765, header_num = 0, file_num = 1, last, line_num = 49}, // 10
// {offset = 5767, header_num = 0, file_num = 1, last, line_num = 50}, // 11
// {offset = 5768, header_num = 0, file_num = 1, last, line_num = 51}, // 12
// {offset = 5773, header_num = 0, file_num = 1, last, line_num = -1}, // 13
// {offset = 5787, header_num = 1, file_num = 1, last, line_num = 19}, // 14
// {offset = 5790, header_num = 1, file_num = 1, last, line_num = 20}, // 15
// {offset = 5793, header_num = 1, file_num = 1, last, line_num = 67}, // 16
// {offset = 5793, header_num = 1, file_num = 1, last, line_num = -1}, // 17
// {offset = 5793, header_num = 1, file_num = 1,!last, line_num = 66}, // 18
// {offset = 5795, header_num = 1, file_num = 1, last, line_num = 68}, // 19
// {offset = 5798, header_num = 1, file_num = 1, last, line_num = -1}, // 20
// The entries with line_num == -1 mark the end of a function: the
// associated offset is one past the last instruction in the
// function. This can correspond to the beginning of the next
// function (as is true for offset 3232); alternately, there can be
// a gap between the end of one function and the start of the next
// (as is true for some others, most obviously from 3236->5764).
//
// Case 1: lookup_key has offset == 10. lower_bound returns
// offsets[0]. Since it's not an exact match and we're
// at the beginning of offsets, we return end() (invalid).
// Case 2: lookup_key has offset 10000. lower_bound returns
// offset[21] (end()). We return end() (invalid).
// Case 3: lookup_key has offset == 3211. lower_bound matches
// offsets[0] exactly, and that's the entry we return.
// Case 4: lookup_key has offset == 3232. lower_bound returns
// offsets[4]. That's an exact match, but indicates
// end-of-function. We check if offsets[5] is also an
// exact match but not end-of-function. It is, so we
// return offsets[5].
// Case 5: lookup_key has offset == 3214. lower_bound returns
// offsets[1]. Since it's not an exact match, we back
// up to the offset that's < lookup_key, offsets[0].
// We note offsets[0] is a valid entry (not end-of-function),
// so that's the entry we return.
// Case 6: lookup_key has offset == 4000. lower_bound returns
// offsets[8]. Since it's not an exact match, we back
// up to offsets[7]. Since offsets[7] indicates
// end-of-function, we know lookup_key is between
// functions, so we return end() (not a valid offset).
// Case 7: lookup_key has offset == 5794. lower_bound returns
// offsets[19]. Since it's not an exact match, we back
// up to offsets[16]. Note we back up to the *first*
// entry with offset 5793, not just offsets[19-1].
// We note offsets[16] is a valid entry, so we return it.
// If offsets[16] had had line_num == -1, we would have
// checked offsets[17]. The reason for this is that
// 16 and 17 can be in an arbitrary order, since we sort
// only by offset and last_line_for_offset. (Note it
// doesn't help to use line_number as a tertiary sort key,
// since sometimes we want the -1 to be first and sometimes
// we want it to be last.)
// This deals with cases (1) and (2).
if ((it == offsets->begin() && offset < it->offset)
|| it == offsets->end())
return offsets->end();
// This deals with cases (3) and (4).
if (offset == it->offset)
{
while (it != offsets->end()
&& it->offset == offset
&& it->line_num == -1)
++it;
if (it == offsets->end() || it->offset != offset)
return offsets->end();
else
return it;
}
// This handles the first part of case (7) -- we back up to the
// *first* entry that has the offset that's behind us.
gold_assert(it != offsets->begin());
std::vector<Offset_to_lineno_entry>::const_iterator range_end = it;
--it;
const off_t range_value = it->offset;
while (it != offsets->begin() && (it-1)->offset == range_value)
--it;
// This handles cases (5), (6), and (7): if any entry in the
// equal_range [it, range_end) has a line_num != -1, it's a valid
// match. If not, we're not in a function. The line number we saw
// last for an offset will be sorted first, so it'll get returned if
// it's present.
for (; it != range_end; ++it)
if (it->line_num != -1)
return it;
return offsets->end();
}
// Returns the canonical filename:lineno for the address passed in.
// If other_lines is not NULL, appends the non-canonical lines
// assigned to the same address.
template<int size, bool big_endian>
std::string
Sized_dwarf_line_info<size, big_endian>::do_addr2line(
unsigned int shndx,
off_t offset,
std::vector<std::string>* other_lines)
{
if (this->data_valid_ == false)
return "";
const std::vector<Offset_to_lineno_entry>* offsets;
// If we do not have reloc information, then our input is a .so or
// some similar data structure where all the information is held in
// the offset. In that case, we ignore the input shndx.
if (this->input_is_relobj())
offsets = &this->line_number_map_[shndx];
else
offsets = &this->line_number_map_[-1U];
if (offsets->empty())
return "";
typename std::vector<Offset_to_lineno_entry>::const_iterator it
= offset_to_iterator(offsets, offset);
if (it == offsets->end())
return "";
std::string result = this->format_file_lineno(*it);
if (other_lines != NULL)
for (++it; it != offsets->end() && it->offset == offset; ++it)
{
if (it->line_num == -1)
continue; // The end of a previous function.
other_lines->push_back(this->format_file_lineno(*it));
}
return result;
}
// Convert the file_num + line_num into a string.
template<int size, bool big_endian>
std::string
Sized_dwarf_line_info<size, big_endian>::format_file_lineno(
const Offset_to_lineno_entry& loc) const
{
std::string ret;
gold_assert(loc.header_num < static_cast<int>(this->files_.size()));
gold_assert(loc.file_num
< static_cast<unsigned int>(this->files_[loc.header_num].size()));
const std::pair<int, std::string>& filename_pair
= this->files_[loc.header_num][loc.file_num];
const std::string& filename = filename_pair.second;
gold_assert(loc.header_num < static_cast<int>(this->directories_.size()));
gold_assert(filename_pair.first
< static_cast<int>(this->directories_[loc.header_num].size()));
const std::string& dirname
= this->directories_[loc.header_num][filename_pair.first];
if (!dirname.empty())
{
ret += dirname;
ret += "/";
}
ret += filename;
if (ret.empty())
ret = "(unknown)";
char buffer[64]; // enough to hold a line number
snprintf(buffer, sizeof(buffer), "%d", loc.line_num);
ret += ":";
ret += buffer;
return ret;
}
// Dwarf_line_info routines.
static unsigned int next_generation_count = 0;
struct Addr2line_cache_entry
{
Object* object;
unsigned int shndx;
Dwarf_line_info* dwarf_line_info;
unsigned int generation_count;
unsigned int access_count;
Addr2line_cache_entry(Object* o, unsigned int s, Dwarf_line_info* d)
: object(o), shndx(s), dwarf_line_info(d),
generation_count(next_generation_count), access_count(0)
{
if (next_generation_count < (1U << 31))
++next_generation_count;
}
};
// We expect this cache to be small, so don't bother with a hashtable
// or priority queue or anything: just use a simple vector.
static std::vector<Addr2line_cache_entry> addr2line_cache;
std::string
Dwarf_line_info::one_addr2line(Object* object,
unsigned int shndx, off_t offset,
size_t cache_size,
std::vector<std::string>* other_lines)
{
Dwarf_line_info* lineinfo = NULL;
std::vector<Addr2line_cache_entry>::iterator it;
// First, check the cache. If we hit, update the counts.
for (it = addr2line_cache.begin(); it != addr2line_cache.end(); ++it)
{
if (it->object == object && it->shndx == shndx)
{
lineinfo = it->dwarf_line_info;
it->generation_count = next_generation_count;
// We cap generation_count at 2^31 -1 to avoid overflow.
if (next_generation_count < (1U << 31))
++next_generation_count;
// We cap access_count at 31 so 2^access_count doesn't overflow
if (it->access_count < 31)
++it->access_count;
break;
}
}
// If we don't hit the cache, create a new object and insert into the
// cache.
if (lineinfo == NULL)
{
switch (parameters->size_and_endianness())
{
#ifdef HAVE_TARGET_32_LITTLE
case Parameters::TARGET_32_LITTLE:
lineinfo = new Sized_dwarf_line_info<32, false>(object, shndx); break;
#endif
#ifdef HAVE_TARGET_32_BIG
case Parameters::TARGET_32_BIG:
lineinfo = new Sized_dwarf_line_info<32, true>(object, shndx); break;
#endif
#ifdef HAVE_TARGET_64_LITTLE
case Parameters::TARGET_64_LITTLE:
lineinfo = new Sized_dwarf_line_info<64, false>(object, shndx); break;
#endif
#ifdef HAVE_TARGET_64_BIG
case Parameters::TARGET_64_BIG:
lineinfo = new Sized_dwarf_line_info<64, true>(object, shndx); break;
#endif
default:
gold_unreachable();
}
addr2line_cache.push_back(Addr2line_cache_entry(object, shndx, lineinfo));
}
// Now that we have our object, figure out the answer
std::string retval = lineinfo->addr2line(shndx, offset, other_lines);
// Finally, if our cache has grown too big, delete old objects. We
// assume the common (probably only) case is deleting only one object.
// We use a pretty simple scheme to evict: function of LRU and MFU.
while (addr2line_cache.size() > cache_size)
{
unsigned int lowest_score = ~0U;
std::vector<Addr2line_cache_entry>::iterator lowest
= addr2line_cache.end();
for (it = addr2line_cache.begin(); it != addr2line_cache.end(); ++it)
{
const unsigned int score = (it->generation_count
+ (1U << it->access_count));
if (score < lowest_score)
{
lowest_score = score;
lowest = it;
}
}
if (lowest != addr2line_cache.end())
{
delete lowest->dwarf_line_info;
addr2line_cache.erase(lowest);
}
}
return retval;
}
void
Dwarf_line_info::clear_addr2line_cache()
{
for (std::vector<Addr2line_cache_entry>::iterator it = addr2line_cache.begin();
it != addr2line_cache.end();
++it)
delete it->dwarf_line_info;
addr2line_cache.clear();
}
#ifdef HAVE_TARGET_32_LITTLE
template
class Sized_dwarf_line_info<32, false>;
#endif
#ifdef HAVE_TARGET_32_BIG
template
class Sized_dwarf_line_info<32, true>;
#endif
#ifdef HAVE_TARGET_64_LITTLE
template
class Sized_dwarf_line_info<64, false>;
#endif
#ifdef HAVE_TARGET_64_BIG
template
class Sized_dwarf_line_info<64, true>;
#endif
} // End namespace gold.
|