aboutsummaryrefslogtreecommitdiff
path: root/gdb/amd64-tdep.c
blob: 08a9cae6c4150aba6e1fa9805ee2e1cfeeca3dc0 (plain)
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
/* Target-dependent code for AMD64.

   Copyright (C) 2001-2017 Free Software Foundation, Inc.

   Contributed by Jiri Smid, SuSE Labs.

   This file is part of GDB.

   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, see <http://www.gnu.org/licenses/>.  */

#include "defs.h"
#include "opcode/i386.h"
#include "dis-asm.h"
#include "arch-utils.h"
#include "block.h"
#include "dummy-frame.h"
#include "frame.h"
#include "frame-base.h"
#include "frame-unwind.h"
#include "inferior.h"
#include "infrun.h"
#include "gdbcmd.h"
#include "gdbcore.h"
#include "objfiles.h"
#include "regcache.h"
#include "regset.h"
#include "symfile.h"
#include "disasm.h"
#include "amd64-tdep.h"
#include "i387-tdep.h"
#include "x86-xstate.h"
#include <algorithm>
#include "target-descriptions.h"
#include "arch/amd64.h"
#include "producer.h"
#include "ax.h"
#include "ax-gdb.h"
#include "common/byte-vector.h"

/* Note that the AMD64 architecture was previously known as x86-64.
   The latter is (forever) engraved into the canonical system name as
   returned by config.guess, and used as the name for the AMD64 port
   of GNU/Linux.  The BSD's have renamed their ports to amd64; they
   don't like to shout.  For GDB we prefer the amd64_-prefix over the
   x86_64_-prefix since it's so much easier to type.  */

/* Register information.  */

static const char *amd64_register_names[] = 
{
  "rax", "rbx", "rcx", "rdx", "rsi", "rdi", "rbp", "rsp",

  /* %r8 is indeed register number 8.  */
  "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15",
  "rip", "eflags", "cs", "ss", "ds", "es", "fs", "gs",

  /* %st0 is register number 24.  */
  "st0", "st1", "st2", "st3", "st4", "st5", "st6", "st7",
  "fctrl", "fstat", "ftag", "fiseg", "fioff", "foseg", "fooff", "fop",

  /* %xmm0 is register number 40.  */
  "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7",
  "xmm8", "xmm9", "xmm10", "xmm11", "xmm12", "xmm13", "xmm14", "xmm15",
  "mxcsr",
};

static const char *amd64_ymm_names[] = 
{
  "ymm0", "ymm1", "ymm2", "ymm3",
  "ymm4", "ymm5", "ymm6", "ymm7",
  "ymm8", "ymm9", "ymm10", "ymm11",
  "ymm12", "ymm13", "ymm14", "ymm15"
};

static const char *amd64_ymm_avx512_names[] =
{
  "ymm16", "ymm17", "ymm18", "ymm19",
  "ymm20", "ymm21", "ymm22", "ymm23",
  "ymm24", "ymm25", "ymm26", "ymm27",
  "ymm28", "ymm29", "ymm30", "ymm31"
};

static const char *amd64_ymmh_names[] = 
{
  "ymm0h", "ymm1h", "ymm2h", "ymm3h",
  "ymm4h", "ymm5h", "ymm6h", "ymm7h",
  "ymm8h", "ymm9h", "ymm10h", "ymm11h",
  "ymm12h", "ymm13h", "ymm14h", "ymm15h"
};

static const char *amd64_ymmh_avx512_names[] =
{
  "ymm16h", "ymm17h", "ymm18h", "ymm19h",
  "ymm20h", "ymm21h", "ymm22h", "ymm23h",
  "ymm24h", "ymm25h", "ymm26h", "ymm27h",
  "ymm28h", "ymm29h", "ymm30h", "ymm31h"
};

static const char *amd64_mpx_names[] =
{
  "bnd0raw", "bnd1raw", "bnd2raw", "bnd3raw", "bndcfgu", "bndstatus"
};

static const char *amd64_k_names[] =
{
  "k0", "k1", "k2", "k3",
  "k4", "k5", "k6", "k7"
};

static const char *amd64_zmmh_names[] =
{
  "zmm0h", "zmm1h", "zmm2h", "zmm3h",
  "zmm4h", "zmm5h", "zmm6h", "zmm7h",
  "zmm8h", "zmm9h", "zmm10h", "zmm11h",
  "zmm12h", "zmm13h", "zmm14h", "zmm15h",
  "zmm16h", "zmm17h", "zmm18h", "zmm19h",
  "zmm20h", "zmm21h", "zmm22h", "zmm23h",
  "zmm24h", "zmm25h", "zmm26h", "zmm27h",
  "zmm28h", "zmm29h", "zmm30h", "zmm31h"
};

static const char *amd64_zmm_names[] =
{
  "zmm0", "zmm1", "zmm2", "zmm3",
  "zmm4", "zmm5", "zmm6", "zmm7",
  "zmm8", "zmm9", "zmm10", "zmm11",
  "zmm12", "zmm13", "zmm14", "zmm15",
  "zmm16", "zmm17", "zmm18", "zmm19",
  "zmm20", "zmm21", "zmm22", "zmm23",
  "zmm24", "zmm25", "zmm26", "zmm27",
  "zmm28", "zmm29", "zmm30", "zmm31"
};

static const char *amd64_xmm_avx512_names[] = {
    "xmm16",  "xmm17",  "xmm18",  "xmm19",
    "xmm20",  "xmm21",  "xmm22",  "xmm23",
    "xmm24",  "xmm25",  "xmm26",  "xmm27",
    "xmm28",  "xmm29",  "xmm30",  "xmm31"
};

static const char *amd64_pkeys_names[] = {
    "pkru"
};

/* DWARF Register Number Mapping as defined in the System V psABI,
   section 3.6.  */

static int amd64_dwarf_regmap[] =
{
  /* General Purpose Registers RAX, RDX, RCX, RBX, RSI, RDI.  */
  AMD64_RAX_REGNUM, AMD64_RDX_REGNUM,
  AMD64_RCX_REGNUM, AMD64_RBX_REGNUM,
  AMD64_RSI_REGNUM, AMD64_RDI_REGNUM,

  /* Frame Pointer Register RBP.  */
  AMD64_RBP_REGNUM,

  /* Stack Pointer Register RSP.  */
  AMD64_RSP_REGNUM,

  /* Extended Integer Registers 8 - 15.  */
  AMD64_R8_REGNUM,		/* %r8 */
  AMD64_R9_REGNUM,		/* %r9 */
  AMD64_R10_REGNUM,		/* %r10 */
  AMD64_R11_REGNUM,		/* %r11 */
  AMD64_R12_REGNUM,		/* %r12 */
  AMD64_R13_REGNUM,		/* %r13 */
  AMD64_R14_REGNUM,		/* %r14 */
  AMD64_R15_REGNUM,		/* %r15 */

  /* Return Address RA.  Mapped to RIP.  */
  AMD64_RIP_REGNUM,

  /* SSE Registers 0 - 7.  */
  AMD64_XMM0_REGNUM + 0, AMD64_XMM1_REGNUM,
  AMD64_XMM0_REGNUM + 2, AMD64_XMM0_REGNUM + 3,
  AMD64_XMM0_REGNUM + 4, AMD64_XMM0_REGNUM + 5,
  AMD64_XMM0_REGNUM + 6, AMD64_XMM0_REGNUM + 7,

  /* Extended SSE Registers 8 - 15.  */
  AMD64_XMM0_REGNUM + 8, AMD64_XMM0_REGNUM + 9,
  AMD64_XMM0_REGNUM + 10, AMD64_XMM0_REGNUM + 11,
  AMD64_XMM0_REGNUM + 12, AMD64_XMM0_REGNUM + 13,
  AMD64_XMM0_REGNUM + 14, AMD64_XMM0_REGNUM + 15,

  /* Floating Point Registers 0-7.  */
  AMD64_ST0_REGNUM + 0, AMD64_ST0_REGNUM + 1,
  AMD64_ST0_REGNUM + 2, AMD64_ST0_REGNUM + 3,
  AMD64_ST0_REGNUM + 4, AMD64_ST0_REGNUM + 5,
  AMD64_ST0_REGNUM + 6, AMD64_ST0_REGNUM + 7,

  /* MMX Registers 0 - 7.
     We have to handle those registers specifically, as their register
     number within GDB depends on the target (or they may even not be
     available at all).  */
  -1, -1, -1, -1, -1, -1, -1, -1,

  /* Control and Status Flags Register.  */
  AMD64_EFLAGS_REGNUM,

  /* Selector Registers.  */
  AMD64_ES_REGNUM,
  AMD64_CS_REGNUM,
  AMD64_SS_REGNUM,
  AMD64_DS_REGNUM,
  AMD64_FS_REGNUM,
  AMD64_GS_REGNUM,
  -1,
  -1,

  /* Segment Base Address Registers.  */
  -1,
  -1,
  -1,
  -1,

  /* Special Selector Registers.  */
  -1,
  -1,

  /* Floating Point Control Registers.  */
  AMD64_MXCSR_REGNUM,
  AMD64_FCTRL_REGNUM,
  AMD64_FSTAT_REGNUM
};

static const int amd64_dwarf_regmap_len =
  (sizeof (amd64_dwarf_regmap) / sizeof (amd64_dwarf_regmap[0]));

/* Convert DWARF register number REG to the appropriate register
   number used by GDB.  */

static int
amd64_dwarf_reg_to_regnum (struct gdbarch *gdbarch, int reg)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  int ymm0_regnum = tdep->ymm0_regnum;
  int regnum = -1;

  if (reg >= 0 && reg < amd64_dwarf_regmap_len)
    regnum = amd64_dwarf_regmap[reg];

  if (ymm0_regnum >= 0
	   && i386_xmm_regnum_p (gdbarch, regnum))
    regnum += ymm0_regnum - I387_XMM0_REGNUM (tdep);

  return regnum;
}

/* Map architectural register numbers to gdb register numbers.  */

static const int amd64_arch_regmap[16] =
{
  AMD64_RAX_REGNUM,	/* %rax */
  AMD64_RCX_REGNUM,	/* %rcx */
  AMD64_RDX_REGNUM,	/* %rdx */
  AMD64_RBX_REGNUM,	/* %rbx */
  AMD64_RSP_REGNUM,	/* %rsp */
  AMD64_RBP_REGNUM,	/* %rbp */
  AMD64_RSI_REGNUM,	/* %rsi */
  AMD64_RDI_REGNUM,	/* %rdi */
  AMD64_R8_REGNUM,	/* %r8 */
  AMD64_R9_REGNUM,	/* %r9 */
  AMD64_R10_REGNUM,	/* %r10 */
  AMD64_R11_REGNUM,	/* %r11 */
  AMD64_R12_REGNUM,	/* %r12 */
  AMD64_R13_REGNUM,	/* %r13 */
  AMD64_R14_REGNUM,	/* %r14 */
  AMD64_R15_REGNUM	/* %r15 */
};

static const int amd64_arch_regmap_len =
  (sizeof (amd64_arch_regmap) / sizeof (amd64_arch_regmap[0]));

/* Convert architectural register number REG to the appropriate register
   number used by GDB.  */

static int
amd64_arch_reg_to_regnum (int reg)
{
  gdb_assert (reg >= 0 && reg < amd64_arch_regmap_len);

  return amd64_arch_regmap[reg];
}

/* Register names for byte pseudo-registers.  */

static const char *amd64_byte_names[] =
{
  "al", "bl", "cl", "dl", "sil", "dil", "bpl", "spl",
  "r8l", "r9l", "r10l", "r11l", "r12l", "r13l", "r14l", "r15l",
  "ah", "bh", "ch", "dh"
};

/* Number of lower byte registers.  */
#define AMD64_NUM_LOWER_BYTE_REGS 16

/* Register names for word pseudo-registers.  */

static const char *amd64_word_names[] =
{
  "ax", "bx", "cx", "dx", "si", "di", "bp", "", 
  "r8w", "r9w", "r10w", "r11w", "r12w", "r13w", "r14w", "r15w"
};

/* Register names for dword pseudo-registers.  */

static const char *amd64_dword_names[] =
{
  "eax", "ebx", "ecx", "edx", "esi", "edi", "ebp", "esp", 
  "r8d", "r9d", "r10d", "r11d", "r12d", "r13d", "r14d", "r15d",
  "eip"
};

/* Return the name of register REGNUM.  */

static const char *
amd64_pseudo_register_name (struct gdbarch *gdbarch, int regnum)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  if (i386_byte_regnum_p (gdbarch, regnum))
    return amd64_byte_names[regnum - tdep->al_regnum];
  else if (i386_zmm_regnum_p (gdbarch, regnum))
    return amd64_zmm_names[regnum - tdep->zmm0_regnum];
  else if (i386_ymm_regnum_p (gdbarch, regnum))
    return amd64_ymm_names[regnum - tdep->ymm0_regnum];
  else if (i386_ymm_avx512_regnum_p (gdbarch, regnum))
    return amd64_ymm_avx512_names[regnum - tdep->ymm16_regnum];
  else if (i386_word_regnum_p (gdbarch, regnum))
    return amd64_word_names[regnum - tdep->ax_regnum];
  else if (i386_dword_regnum_p (gdbarch, regnum))
    return amd64_dword_names[regnum - tdep->eax_regnum];
  else
    return i386_pseudo_register_name (gdbarch, regnum);
}

static struct value *
amd64_pseudo_register_read_value (struct gdbarch *gdbarch,
				  struct regcache *regcache,
				  int regnum)
{
  gdb_byte *raw_buf = (gdb_byte *) alloca (register_size (gdbarch, regnum));
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  enum register_status status;
  struct value *result_value;
  gdb_byte *buf;

  result_value = allocate_value (register_type (gdbarch, regnum));
  VALUE_LVAL (result_value) = lval_register;
  VALUE_REGNUM (result_value) = regnum;
  buf = value_contents_raw (result_value);

  if (i386_byte_regnum_p (gdbarch, regnum))
    {
      int gpnum = regnum - tdep->al_regnum;

      /* Extract (always little endian).  */
      if (gpnum >= AMD64_NUM_LOWER_BYTE_REGS)
	{
	  /* Special handling for AH, BH, CH, DH.  */
	  status = regcache_raw_read (regcache,
				      gpnum - AMD64_NUM_LOWER_BYTE_REGS,
				      raw_buf);
	  if (status == REG_VALID)
	    memcpy (buf, raw_buf + 1, 1);
	  else
	    mark_value_bytes_unavailable (result_value, 0,
					  TYPE_LENGTH (value_type (result_value)));
	}
      else
	{
	  status = regcache_raw_read (regcache, gpnum, raw_buf);
	  if (status == REG_VALID)
	    memcpy (buf, raw_buf, 1);
	  else
	    mark_value_bytes_unavailable (result_value, 0,
					  TYPE_LENGTH (value_type (result_value)));
	}
    }
  else if (i386_dword_regnum_p (gdbarch, regnum))
    {
      int gpnum = regnum - tdep->eax_regnum;
      /* Extract (always little endian).  */
      status = regcache_raw_read (regcache, gpnum, raw_buf);
      if (status == REG_VALID)
	memcpy (buf, raw_buf, 4);
      else
	mark_value_bytes_unavailable (result_value, 0,
				      TYPE_LENGTH (value_type (result_value)));
    }
  else
    i386_pseudo_register_read_into_value (gdbarch, regcache, regnum,
					  result_value);

  return result_value;
}

static void
amd64_pseudo_register_write (struct gdbarch *gdbarch,
			     struct regcache *regcache,
			     int regnum, const gdb_byte *buf)
{
  gdb_byte *raw_buf = (gdb_byte *) alloca (register_size (gdbarch, regnum));
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  if (i386_byte_regnum_p (gdbarch, regnum))
    {
      int gpnum = regnum - tdep->al_regnum;

      if (gpnum >= AMD64_NUM_LOWER_BYTE_REGS)
	{
	  /* Read ... AH, BH, CH, DH.  */
	  regcache_raw_read (regcache,
			     gpnum - AMD64_NUM_LOWER_BYTE_REGS, raw_buf);
	  /* ... Modify ... (always little endian).  */
	  memcpy (raw_buf + 1, buf, 1);
	  /* ... Write.  */
	  regcache_raw_write (regcache,
			      gpnum - AMD64_NUM_LOWER_BYTE_REGS, raw_buf);
	}
      else
	{
	  /* Read ...  */
	  regcache_raw_read (regcache, gpnum, raw_buf);
	  /* ... Modify ... (always little endian).  */
	  memcpy (raw_buf, buf, 1);
	  /* ... Write.  */
	  regcache_raw_write (regcache, gpnum, raw_buf);
	}
    }
  else if (i386_dword_regnum_p (gdbarch, regnum))
    {
      int gpnum = regnum - tdep->eax_regnum;

      /* Read ...  */
      regcache_raw_read (regcache, gpnum, raw_buf);
      /* ... Modify ... (always little endian).  */
      memcpy (raw_buf, buf, 4);
      /* ... Write.  */
      regcache_raw_write (regcache, gpnum, raw_buf);
    }
  else
    i386_pseudo_register_write (gdbarch, regcache, regnum, buf);
}

/* Implement the 'ax_pseudo_register_collect' gdbarch method.  */

static int
amd64_ax_pseudo_register_collect (struct gdbarch *gdbarch,
				  struct agent_expr *ax, int regnum)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  if (i386_byte_regnum_p (gdbarch, regnum))
    {
      int gpnum = regnum - tdep->al_regnum;

      if (gpnum >= AMD64_NUM_LOWER_BYTE_REGS)
	ax_reg_mask (ax, gpnum - AMD64_NUM_LOWER_BYTE_REGS);
      else
	ax_reg_mask (ax, gpnum);
      return 0;
    }
  else if (i386_dword_regnum_p (gdbarch, regnum))
    {
      int gpnum = regnum - tdep->eax_regnum;

      ax_reg_mask (ax, gpnum);
      return 0;
    }
  else
    return i386_ax_pseudo_register_collect (gdbarch, ax, regnum);
}



/* Register classes as defined in the psABI.  */

enum amd64_reg_class
{
  AMD64_INTEGER,
  AMD64_SSE,
  AMD64_SSEUP,
  AMD64_X87,
  AMD64_X87UP,
  AMD64_COMPLEX_X87,
  AMD64_NO_CLASS,
  AMD64_MEMORY
};

/* Return the union class of CLASS1 and CLASS2.  See the psABI for
   details.  */

static enum amd64_reg_class
amd64_merge_classes (enum amd64_reg_class class1, enum amd64_reg_class class2)
{
  /* Rule (a): If both classes are equal, this is the resulting class.  */
  if (class1 == class2)
    return class1;

  /* Rule (b): If one of the classes is NO_CLASS, the resulting class
     is the other class.  */
  if (class1 == AMD64_NO_CLASS)
    return class2;
  if (class2 == AMD64_NO_CLASS)
    return class1;

  /* Rule (c): If one of the classes is MEMORY, the result is MEMORY.  */
  if (class1 == AMD64_MEMORY || class2 == AMD64_MEMORY)
    return AMD64_MEMORY;

  /* Rule (d): If one of the classes is INTEGER, the result is INTEGER.  */
  if (class1 == AMD64_INTEGER || class2 == AMD64_INTEGER)
    return AMD64_INTEGER;

  /* Rule (e): If one of the classes is X87, X87UP, COMPLEX_X87 class,
     MEMORY is used as class.  */
  if (class1 == AMD64_X87 || class1 == AMD64_X87UP
      || class1 == AMD64_COMPLEX_X87 || class2 == AMD64_X87
      || class2 == AMD64_X87UP || class2 == AMD64_COMPLEX_X87)
    return AMD64_MEMORY;

  /* Rule (f): Otherwise class SSE is used.  */
  return AMD64_SSE;
}

static void amd64_classify (struct type *type, enum amd64_reg_class theclass[2]);

/* Return non-zero if TYPE is a non-POD structure or union type.  */

static int
amd64_non_pod_p (struct type *type)
{
  /* ??? A class with a base class certainly isn't POD, but does this
     catch all non-POD structure types?  */
  if (TYPE_CODE (type) == TYPE_CODE_STRUCT && TYPE_N_BASECLASSES (type) > 0)
    return 1;

  return 0;
}

/* Classify TYPE according to the rules for aggregate (structures and
   arrays) and union types, and store the result in CLASS.  */

static void
amd64_classify_aggregate (struct type *type, enum amd64_reg_class theclass[2])
{
  /* 1. If the size of an object is larger than two eightbytes, or in
        C++, is a non-POD structure or union type, or contains
        unaligned fields, it has class memory.  */
  if (TYPE_LENGTH (type) > 16 || amd64_non_pod_p (type))
    {
      theclass[0] = theclass[1] = AMD64_MEMORY;
      return;
    }

  /* 2. Both eightbytes get initialized to class NO_CLASS.  */
  theclass[0] = theclass[1] = AMD64_NO_CLASS;

  /* 3. Each field of an object is classified recursively so that
        always two fields are considered. The resulting class is
        calculated according to the classes of the fields in the
        eightbyte: */

  if (TYPE_CODE (type) == TYPE_CODE_ARRAY)
    {
      struct type *subtype = check_typedef (TYPE_TARGET_TYPE (type));

      /* All fields in an array have the same type.  */
      amd64_classify (subtype, theclass);
      if (TYPE_LENGTH (type) > 8 && theclass[1] == AMD64_NO_CLASS)
	theclass[1] = theclass[0];
    }
  else
    {
      int i;

      /* Structure or union.  */
      gdb_assert (TYPE_CODE (type) == TYPE_CODE_STRUCT
		  || TYPE_CODE (type) == TYPE_CODE_UNION);

      for (i = 0; i < TYPE_NFIELDS (type); i++)
	{
	  struct type *subtype = check_typedef (TYPE_FIELD_TYPE (type, i));
	  int pos = TYPE_FIELD_BITPOS (type, i) / 64;
	  enum amd64_reg_class subclass[2];
	  int bitsize = TYPE_FIELD_BITSIZE (type, i);
	  int endpos;

	  if (bitsize == 0)
	    bitsize = TYPE_LENGTH (subtype) * 8;
	  endpos = (TYPE_FIELD_BITPOS (type, i) + bitsize - 1) / 64;

	  /* Ignore static fields.  */
	  if (field_is_static (&TYPE_FIELD (type, i)))
	    continue;

	  gdb_assert (pos == 0 || pos == 1);

	  amd64_classify (subtype, subclass);
	  theclass[pos] = amd64_merge_classes (theclass[pos], subclass[0]);
	  if (bitsize <= 64 && pos == 0 && endpos == 1)
	    /* This is a bit of an odd case:  We have a field that would
	       normally fit in one of the two eightbytes, except that
	       it is placed in a way that this field straddles them.
	       This has been seen with a structure containing an array.

	       The ABI is a bit unclear in this case, but we assume that
	       this field's class (stored in subclass[0]) must also be merged
	       into class[1].  In other words, our field has a piece stored
	       in the second eight-byte, and thus its class applies to
	       the second eight-byte as well.

	       In the case where the field length exceeds 8 bytes,
	       it should not be necessary to merge the field class
	       into class[1].  As LEN > 8, subclass[1] is necessarily
	       different from AMD64_NO_CLASS.  If subclass[1] is equal
	       to subclass[0], then the normal class[1]/subclass[1]
	       merging will take care of everything.  For subclass[1]
	       to be different from subclass[0], I can only see the case
	       where we have a SSE/SSEUP or X87/X87UP pair, which both
	       use up all 16 bytes of the aggregate, and are already
	       handled just fine (because each portion sits on its own
	       8-byte).  */
	    theclass[1] = amd64_merge_classes (theclass[1], subclass[0]);
	  if (pos == 0)
	    theclass[1] = amd64_merge_classes (theclass[1], subclass[1]);
	}
    }

  /* 4. Then a post merger cleanup is done:  */

  /* Rule (a): If one of the classes is MEMORY, the whole argument is
     passed in memory.  */
  if (theclass[0] == AMD64_MEMORY || theclass[1] == AMD64_MEMORY)
    theclass[0] = theclass[1] = AMD64_MEMORY;

  /* Rule (b): If SSEUP is not preceded by SSE, it is converted to
     SSE.  */
  if (theclass[0] == AMD64_SSEUP)
    theclass[0] = AMD64_SSE;
  if (theclass[1] == AMD64_SSEUP && theclass[0] != AMD64_SSE)
    theclass[1] = AMD64_SSE;
}

/* Classify TYPE, and store the result in CLASS.  */

static void
amd64_classify (struct type *type, enum amd64_reg_class theclass[2])
{
  enum type_code code = TYPE_CODE (type);
  int len = TYPE_LENGTH (type);

  theclass[0] = theclass[1] = AMD64_NO_CLASS;

  /* Arguments of types (signed and unsigned) _Bool, char, short, int,
     long, long long, and pointers are in the INTEGER class.  Similarly,
     range types, used by languages such as Ada, are also in the INTEGER
     class.  */
  if ((code == TYPE_CODE_INT || code == TYPE_CODE_ENUM
       || code == TYPE_CODE_BOOL || code == TYPE_CODE_RANGE
       || code == TYPE_CODE_CHAR
       || code == TYPE_CODE_PTR || TYPE_IS_REFERENCE (type))
      && (len == 1 || len == 2 || len == 4 || len == 8))
    theclass[0] = AMD64_INTEGER;

  /* Arguments of types float, double, _Decimal32, _Decimal64 and __m64
     are in class SSE.  */
  else if ((code == TYPE_CODE_FLT || code == TYPE_CODE_DECFLOAT)
	   && (len == 4 || len == 8))
    /* FIXME: __m64 .  */
    theclass[0] = AMD64_SSE;

  /* Arguments of types __float128, _Decimal128 and __m128 are split into
     two halves.  The least significant ones belong to class SSE, the most
     significant one to class SSEUP.  */
  else if (code == TYPE_CODE_DECFLOAT && len == 16)
    /* FIXME: __float128, __m128.  */
    theclass[0] = AMD64_SSE, theclass[1] = AMD64_SSEUP;

  /* The 64-bit mantissa of arguments of type long double belongs to
     class X87, the 16-bit exponent plus 6 bytes of padding belongs to
     class X87UP.  */
  else if (code == TYPE_CODE_FLT && len == 16)
    /* Class X87 and X87UP.  */
    theclass[0] = AMD64_X87, theclass[1] = AMD64_X87UP;

  /* Arguments of complex T where T is one of the types float or
     double get treated as if they are implemented as:

     struct complexT {
       T real;
       T imag;
     };

  */
  else if (code == TYPE_CODE_COMPLEX && len == 8)
    theclass[0] = AMD64_SSE;
  else if (code == TYPE_CODE_COMPLEX && len == 16)
    theclass[0] = theclass[1] = AMD64_SSE;

  /* A variable of type complex long double is classified as type
     COMPLEX_X87.  */
  else if (code == TYPE_CODE_COMPLEX && len == 32)
    theclass[0] = AMD64_COMPLEX_X87;

  /* Aggregates.  */
  else if (code == TYPE_CODE_ARRAY || code == TYPE_CODE_STRUCT
	   || code == TYPE_CODE_UNION)
    amd64_classify_aggregate (type, theclass);
}

static enum return_value_convention
amd64_return_value (struct gdbarch *gdbarch, struct value *function,
		    struct type *type, struct regcache *regcache,
		    gdb_byte *readbuf, const gdb_byte *writebuf)
{
  enum amd64_reg_class theclass[2];
  int len = TYPE_LENGTH (type);
  static int integer_regnum[] = { AMD64_RAX_REGNUM, AMD64_RDX_REGNUM };
  static int sse_regnum[] = { AMD64_XMM0_REGNUM, AMD64_XMM1_REGNUM };
  int integer_reg = 0;
  int sse_reg = 0;
  int i;

  gdb_assert (!(readbuf && writebuf));

  /* 1. Classify the return type with the classification algorithm.  */
  amd64_classify (type, theclass);

  /* 2. If the type has class MEMORY, then the caller provides space
     for the return value and passes the address of this storage in
     %rdi as if it were the first argument to the function.  In effect,
     this address becomes a hidden first argument.

     On return %rax will contain the address that has been passed in
     by the caller in %rdi.  */
  if (theclass[0] == AMD64_MEMORY)
    {
      /* As indicated by the comment above, the ABI guarantees that we
         can always find the return value just after the function has
         returned.  */

      if (readbuf)
	{
	  ULONGEST addr;

	  regcache_raw_read_unsigned (regcache, AMD64_RAX_REGNUM, &addr);
	  read_memory (addr, readbuf, TYPE_LENGTH (type));
	}

      return RETURN_VALUE_ABI_RETURNS_ADDRESS;
    }

  /* 8. If the class is COMPLEX_X87, the real part of the value is
        returned in %st0 and the imaginary part in %st1.  */
  if (theclass[0] == AMD64_COMPLEX_X87)
    {
      if (readbuf)
	{
	  regcache_raw_read (regcache, AMD64_ST0_REGNUM, readbuf);
	  regcache_raw_read (regcache, AMD64_ST1_REGNUM, readbuf + 16);
	}

      if (writebuf)
	{
	  i387_return_value (gdbarch, regcache);
	  regcache_raw_write (regcache, AMD64_ST0_REGNUM, writebuf);
	  regcache_raw_write (regcache, AMD64_ST1_REGNUM, writebuf + 16);

	  /* Fix up the tag word such that both %st(0) and %st(1) are
	     marked as valid.  */
	  regcache_raw_write_unsigned (regcache, AMD64_FTAG_REGNUM, 0xfff);
	}

      return RETURN_VALUE_REGISTER_CONVENTION;
    }

  gdb_assert (theclass[1] != AMD64_MEMORY);
  gdb_assert (len <= 16);

  for (i = 0; len > 0; i++, len -= 8)
    {
      int regnum = -1;
      int offset = 0;

      switch (theclass[i])
	{
	case AMD64_INTEGER:
	  /* 3. If the class is INTEGER, the next available register
	     of the sequence %rax, %rdx is used.  */
	  regnum = integer_regnum[integer_reg++];
	  break;

	case AMD64_SSE:
	  /* 4. If the class is SSE, the next available SSE register
             of the sequence %xmm0, %xmm1 is used.  */
	  regnum = sse_regnum[sse_reg++];
	  break;

	case AMD64_SSEUP:
	  /* 5. If the class is SSEUP, the eightbyte is passed in the
	     upper half of the last used SSE register.  */
	  gdb_assert (sse_reg > 0);
	  regnum = sse_regnum[sse_reg - 1];
	  offset = 8;
	  break;

	case AMD64_X87:
	  /* 6. If the class is X87, the value is returned on the X87
             stack in %st0 as 80-bit x87 number.  */
	  regnum = AMD64_ST0_REGNUM;
	  if (writebuf)
	    i387_return_value (gdbarch, regcache);
	  break;

	case AMD64_X87UP:
	  /* 7. If the class is X87UP, the value is returned together
             with the previous X87 value in %st0.  */
	  gdb_assert (i > 0 && theclass[0] == AMD64_X87);
	  regnum = AMD64_ST0_REGNUM;
	  offset = 8;
	  len = 2;
	  break;

	case AMD64_NO_CLASS:
	  continue;

	default:
	  gdb_assert (!"Unexpected register class.");
	}

      gdb_assert (regnum != -1);

      if (readbuf)
	regcache_raw_read_part (regcache, regnum, offset, std::min (len, 8),
				readbuf + i * 8);
      if (writebuf)
	regcache_raw_write_part (regcache, regnum, offset, std::min (len, 8),
				 writebuf + i * 8);
    }

  return RETURN_VALUE_REGISTER_CONVENTION;
}


static CORE_ADDR
amd64_push_arguments (struct regcache *regcache, int nargs,
		      struct value **args, CORE_ADDR sp, int struct_return)
{
  static int integer_regnum[] =
  {
    AMD64_RDI_REGNUM,		/* %rdi */
    AMD64_RSI_REGNUM,		/* %rsi */
    AMD64_RDX_REGNUM,		/* %rdx */
    AMD64_RCX_REGNUM,		/* %rcx */
    AMD64_R8_REGNUM,		/* %r8 */
    AMD64_R9_REGNUM		/* %r9 */
  };
  static int sse_regnum[] =
  {
    /* %xmm0 ... %xmm7 */
    AMD64_XMM0_REGNUM + 0, AMD64_XMM1_REGNUM,
    AMD64_XMM0_REGNUM + 2, AMD64_XMM0_REGNUM + 3,
    AMD64_XMM0_REGNUM + 4, AMD64_XMM0_REGNUM + 5,
    AMD64_XMM0_REGNUM + 6, AMD64_XMM0_REGNUM + 7,
  };
  struct value **stack_args = XALLOCAVEC (struct value *, nargs);
  int num_stack_args = 0;
  int num_elements = 0;
  int element = 0;
  int integer_reg = 0;
  int sse_reg = 0;
  int i;

  /* Reserve a register for the "hidden" argument.  */
  if (struct_return)
    integer_reg++;

  for (i = 0; i < nargs; i++)
    {
      struct type *type = value_type (args[i]);
      int len = TYPE_LENGTH (type);
      enum amd64_reg_class theclass[2];
      int needed_integer_regs = 0;
      int needed_sse_regs = 0;
      int j;

      /* Classify argument.  */
      amd64_classify (type, theclass);

      /* Calculate the number of integer and SSE registers needed for
         this argument.  */
      for (j = 0; j < 2; j++)
	{
	  if (theclass[j] == AMD64_INTEGER)
	    needed_integer_regs++;
	  else if (theclass[j] == AMD64_SSE)
	    needed_sse_regs++;
	}

      /* Check whether enough registers are available, and if the
         argument should be passed in registers at all.  */
      if (integer_reg + needed_integer_regs > ARRAY_SIZE (integer_regnum)
	  || sse_reg + needed_sse_regs > ARRAY_SIZE (sse_regnum)
	  || (needed_integer_regs == 0 && needed_sse_regs == 0))
	{
	  /* The argument will be passed on the stack.  */
	  num_elements += ((len + 7) / 8);
	  stack_args[num_stack_args++] = args[i];
	}
      else
	{
	  /* The argument will be passed in registers.  */
	  const gdb_byte *valbuf = value_contents (args[i]);
	  gdb_byte buf[8];

	  gdb_assert (len <= 16);

	  for (j = 0; len > 0; j++, len -= 8)
	    {
	      int regnum = -1;
	      int offset = 0;

	      switch (theclass[j])
		{
		case AMD64_INTEGER:
		  regnum = integer_regnum[integer_reg++];
		  break;

		case AMD64_SSE:
		  regnum = sse_regnum[sse_reg++];
		  break;

		case AMD64_SSEUP:
		  gdb_assert (sse_reg > 0);
		  regnum = sse_regnum[sse_reg - 1];
		  offset = 8;
		  break;

		default:
		  gdb_assert (!"Unexpected register class.");
		}

	      gdb_assert (regnum != -1);
	      memset (buf, 0, sizeof buf);
	      memcpy (buf, valbuf + j * 8, std::min (len, 8));
	      regcache_raw_write_part (regcache, regnum, offset, 8, buf);
	    }
	}
    }

  /* Allocate space for the arguments on the stack.  */
  sp -= num_elements * 8;

  /* The psABI says that "The end of the input argument area shall be
     aligned on a 16 byte boundary."  */
  sp &= ~0xf;

  /* Write out the arguments to the stack.  */
  for (i = 0; i < num_stack_args; i++)
    {
      struct type *type = value_type (stack_args[i]);
      const gdb_byte *valbuf = value_contents (stack_args[i]);
      int len = TYPE_LENGTH (type);

      write_memory (sp + element * 8, valbuf, len);
      element += ((len + 7) / 8);
    }

  /* The psABI says that "For calls that may call functions that use
     varargs or stdargs (prototype-less calls or calls to functions
     containing ellipsis (...) in the declaration) %al is used as
     hidden argument to specify the number of SSE registers used.  */
  regcache_raw_write_unsigned (regcache, AMD64_RAX_REGNUM, sse_reg);
  return sp; 
}

static CORE_ADDR
amd64_push_dummy_call (struct gdbarch *gdbarch, struct value *function,
		       struct regcache *regcache, CORE_ADDR bp_addr,
		       int nargs, struct value **args,	CORE_ADDR sp,
		       int struct_return, CORE_ADDR struct_addr)
{
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  gdb_byte buf[8];

  /* BND registers can be in arbitrary values at the moment of the
     inferior call.  This can cause boundary violations that are not
     due to a real bug or even desired by the user.  The best to be done
     is set the BND registers to allow access to the whole memory, INIT
     state, before pushing the inferior call.   */
  i387_reset_bnd_regs (gdbarch, regcache);

  /* Pass arguments.  */
  sp = amd64_push_arguments (regcache, nargs, args, sp, struct_return);

  /* Pass "hidden" argument".  */
  if (struct_return)
    {
      store_unsigned_integer (buf, 8, byte_order, struct_addr);
      regcache_cooked_write (regcache, AMD64_RDI_REGNUM, buf);
    }

  /* Store return address.  */
  sp -= 8;
  store_unsigned_integer (buf, 8, byte_order, bp_addr);
  write_memory (sp, buf, 8);

  /* Finally, update the stack pointer...  */
  store_unsigned_integer (buf, 8, byte_order, sp);
  regcache_cooked_write (regcache, AMD64_RSP_REGNUM, buf);

  /* ...and fake a frame pointer.  */
  regcache_cooked_write (regcache, AMD64_RBP_REGNUM, buf);

  return sp + 16;
}

/* Displaced instruction handling.  */

/* A partially decoded instruction.
   This contains enough details for displaced stepping purposes.  */

struct amd64_insn
{
  /* The number of opcode bytes.  */
  int opcode_len;
  /* The offset of the rex prefix or -1 if not present.  */
  int rex_offset;
  /* The offset to the first opcode byte.  */
  int opcode_offset;
  /* The offset to the modrm byte or -1 if not present.  */
  int modrm_offset;

  /* The raw instruction.  */
  gdb_byte *raw_insn;
};

struct amd64_displaced_step_closure : public displaced_step_closure
{
  amd64_displaced_step_closure (int insn_buf_len)
  : insn_buf (insn_buf_len, 0)
  {}

  /* For rip-relative insns, saved copy of the reg we use instead of %rip.  */
  int tmp_used = 0;
  int tmp_regno;
  ULONGEST tmp_save;

  /* Details of the instruction.  */
  struct amd64_insn insn_details;

  /* The possibly modified insn.  */
  gdb::byte_vector insn_buf;
};

/* WARNING: Keep onebyte_has_modrm, twobyte_has_modrm in sync with
   ../opcodes/i386-dis.c (until libopcodes exports them, or an alternative,
   at which point delete these in favor of libopcodes' versions).  */

static const unsigned char onebyte_has_modrm[256] = {
  /*	   0 1 2 3 4 5 6 7 8 9 a b c d e f	  */
  /*	   -------------------------------	  */
  /* 00 */ 1,1,1,1,0,0,0,0,1,1,1,1,0,0,0,0, /* 00 */
  /* 10 */ 1,1,1,1,0,0,0,0,1,1,1,1,0,0,0,0, /* 10 */
  /* 20 */ 1,1,1,1,0,0,0,0,1,1,1,1,0,0,0,0, /* 20 */
  /* 30 */ 1,1,1,1,0,0,0,0,1,1,1,1,0,0,0,0, /* 30 */
  /* 40 */ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 40 */
  /* 50 */ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 50 */
  /* 60 */ 0,0,1,1,0,0,0,0,0,1,0,1,0,0,0,0, /* 60 */
  /* 70 */ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 70 */
  /* 80 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* 80 */
  /* 90 */ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 90 */
  /* a0 */ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* a0 */
  /* b0 */ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* b0 */
  /* c0 */ 1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0, /* c0 */
  /* d0 */ 1,1,1,1,0,0,0,0,1,1,1,1,1,1,1,1, /* d0 */
  /* e0 */ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* e0 */
  /* f0 */ 0,0,0,0,0,0,1,1,0,0,0,0,0,0,1,1  /* f0 */
  /*	   -------------------------------	  */
  /*	   0 1 2 3 4 5 6 7 8 9 a b c d e f	  */
};

static const unsigned char twobyte_has_modrm[256] = {
  /*	   0 1 2 3 4 5 6 7 8 9 a b c d e f	  */
  /*	   -------------------------------	  */
  /* 00 */ 1,1,1,1,0,0,0,0,0,0,0,0,0,1,0,1, /* 0f */
  /* 10 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* 1f */
  /* 20 */ 1,1,1,1,1,1,1,0,1,1,1,1,1,1,1,1, /* 2f */
  /* 30 */ 0,0,0,0,0,0,0,0,1,0,1,0,0,0,0,0, /* 3f */
  /* 40 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* 4f */
  /* 50 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* 5f */
  /* 60 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* 6f */
  /* 70 */ 1,1,1,1,1,1,1,0,1,1,1,1,1,1,1,1, /* 7f */
  /* 80 */ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 8f */
  /* 90 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* 9f */
  /* a0 */ 0,0,0,1,1,1,1,1,0,0,0,1,1,1,1,1, /* af */
  /* b0 */ 1,1,1,1,1,1,1,1,1,0,1,1,1,1,1,1, /* bf */
  /* c0 */ 1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0, /* cf */
  /* d0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* df */
  /* e0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ef */
  /* f0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0  /* ff */
  /*	   -------------------------------	  */
  /*	   0 1 2 3 4 5 6 7 8 9 a b c d e f	  */
};

static int amd64_syscall_p (const struct amd64_insn *insn, int *lengthp);

static int
rex_prefix_p (gdb_byte pfx)
{
  return REX_PREFIX_P (pfx);
}

/* Skip the legacy instruction prefixes in INSN.
   We assume INSN is properly sentineled so we don't have to worry
   about falling off the end of the buffer.  */

static gdb_byte *
amd64_skip_prefixes (gdb_byte *insn)
{
  while (1)
    {
      switch (*insn)
	{
	case DATA_PREFIX_OPCODE:
	case ADDR_PREFIX_OPCODE:
	case CS_PREFIX_OPCODE:
	case DS_PREFIX_OPCODE:
	case ES_PREFIX_OPCODE:
	case FS_PREFIX_OPCODE:
	case GS_PREFIX_OPCODE:
	case SS_PREFIX_OPCODE:
	case LOCK_PREFIX_OPCODE:
	case REPE_PREFIX_OPCODE:
	case REPNE_PREFIX_OPCODE:
	  ++insn;
	  continue;
	default:
	  break;
	}
      break;
    }

  return insn;
}

/* Return an integer register (other than RSP) that is unused as an input
   operand in INSN.
   In order to not require adding a rex prefix if the insn doesn't already
   have one, the result is restricted to RAX ... RDI, sans RSP.
   The register numbering of the result follows architecture ordering,
   e.g. RDI = 7.  */

static int
amd64_get_unused_input_int_reg (const struct amd64_insn *details)
{
  /* 1 bit for each reg */
  int used_regs_mask = 0;

  /* There can be at most 3 int regs used as inputs in an insn, and we have
     7 to choose from (RAX ... RDI, sans RSP).
     This allows us to take a conservative approach and keep things simple.
     E.g. By avoiding RAX, we don't have to specifically watch for opcodes
     that implicitly specify RAX.  */

  /* Avoid RAX.  */
  used_regs_mask |= 1 << EAX_REG_NUM;
  /* Similarily avoid RDX, implicit operand in divides.  */
  used_regs_mask |= 1 << EDX_REG_NUM;
  /* Avoid RSP.  */
  used_regs_mask |= 1 << ESP_REG_NUM;

  /* If the opcode is one byte long and there's no ModRM byte,
     assume the opcode specifies a register.  */
  if (details->opcode_len == 1 && details->modrm_offset == -1)
    used_regs_mask |= 1 << (details->raw_insn[details->opcode_offset] & 7);

  /* Mark used regs in the modrm/sib bytes.  */
  if (details->modrm_offset != -1)
    {
      int modrm = details->raw_insn[details->modrm_offset];
      int mod = MODRM_MOD_FIELD (modrm);
      int reg = MODRM_REG_FIELD (modrm);
      int rm = MODRM_RM_FIELD (modrm);
      int have_sib = mod != 3 && rm == 4;

      /* Assume the reg field of the modrm byte specifies a register.  */
      used_regs_mask |= 1 << reg;

      if (have_sib)
	{
	  int base = SIB_BASE_FIELD (details->raw_insn[details->modrm_offset + 1]);
	  int idx = SIB_INDEX_FIELD (details->raw_insn[details->modrm_offset + 1]);
	  used_regs_mask |= 1 << base;
	  used_regs_mask |= 1 << idx;
	}
      else
	{
	  used_regs_mask |= 1 << rm;
	}
    }

  gdb_assert (used_regs_mask < 256);
  gdb_assert (used_regs_mask != 255);

  /* Finally, find a free reg.  */
  {
    int i;

    for (i = 0; i < 8; ++i)
      {
	if (! (used_regs_mask & (1 << i)))
	  return i;
      }

    /* We shouldn't get here.  */
    internal_error (__FILE__, __LINE__, _("unable to find free reg"));
  }
}

/* Extract the details of INSN that we need.  */

static void
amd64_get_insn_details (gdb_byte *insn, struct amd64_insn *details)
{
  gdb_byte *start = insn;
  int need_modrm;

  details->raw_insn = insn;

  details->opcode_len = -1;
  details->rex_offset = -1;
  details->opcode_offset = -1;
  details->modrm_offset = -1;

  /* Skip legacy instruction prefixes.  */
  insn = amd64_skip_prefixes (insn);

  /* Skip REX instruction prefix.  */
  if (rex_prefix_p (*insn))
    {
      details->rex_offset = insn - start;
      ++insn;
    }

  details->opcode_offset = insn - start;

  if (*insn == TWO_BYTE_OPCODE_ESCAPE)
    {
      /* Two or three-byte opcode.  */
      ++insn;
      need_modrm = twobyte_has_modrm[*insn];

      /* Check for three-byte opcode.  */
      switch (*insn)
	{
	case 0x24:
	case 0x25:
	case 0x38:
	case 0x3a:
	case 0x7a:
	case 0x7b:
	  ++insn;
	  details->opcode_len = 3;
	  break;
	default:
	  details->opcode_len = 2;
	  break;
	}
    }
  else
    {
      /* One-byte opcode.  */
      need_modrm = onebyte_has_modrm[*insn];
      details->opcode_len = 1;
    }

  if (need_modrm)
    {
      ++insn;
      details->modrm_offset = insn - start;
    }
}

/* Update %rip-relative addressing in INSN.

   %rip-relative addressing only uses a 32-bit displacement.
   32 bits is not enough to be guaranteed to cover the distance between where
   the real instruction is and where its copy is.
   Convert the insn to use base+disp addressing.
   We set base = pc + insn_length so we can leave disp unchanged.  */

static void
fixup_riprel (struct gdbarch *gdbarch, amd64_displaced_step_closure *dsc,
	      CORE_ADDR from, CORE_ADDR to, struct regcache *regs)
{
  const struct amd64_insn *insn_details = &dsc->insn_details;
  int modrm_offset = insn_details->modrm_offset;
  gdb_byte *insn = insn_details->raw_insn + modrm_offset;
  CORE_ADDR rip_base;
  int insn_length;
  int arch_tmp_regno, tmp_regno;
  ULONGEST orig_value;

  /* %rip+disp32 addressing mode, displacement follows ModRM byte.  */
  ++insn;

  /* Compute the rip-relative address.	*/
  insn_length = gdb_buffered_insn_length (gdbarch, dsc->insn_buf.data (),
					  dsc->insn_buf.size (), from);
  rip_base = from + insn_length;

  /* We need a register to hold the address.
     Pick one not used in the insn.
     NOTE: arch_tmp_regno uses architecture ordering, e.g. RDI = 7.  */
  arch_tmp_regno = amd64_get_unused_input_int_reg (insn_details);
  tmp_regno = amd64_arch_reg_to_regnum (arch_tmp_regno);

  /* REX.B should be unset as we were using rip-relative addressing,
     but ensure it's unset anyway, tmp_regno is not r8-r15.  */
  if (insn_details->rex_offset != -1)
    dsc->insn_buf[insn_details->rex_offset] &= ~REX_B;

  regcache_cooked_read_unsigned (regs, tmp_regno, &orig_value);
  dsc->tmp_regno = tmp_regno;
  dsc->tmp_save = orig_value;
  dsc->tmp_used = 1;

  /* Convert the ModRM field to be base+disp.  */
  dsc->insn_buf[modrm_offset] &= ~0xc7;
  dsc->insn_buf[modrm_offset] |= 0x80 + arch_tmp_regno;

  regcache_cooked_write_unsigned (regs, tmp_regno, rip_base);

  if (debug_displaced)
    fprintf_unfiltered (gdb_stdlog, "displaced: %%rip-relative addressing used.\n"
			"displaced: using temp reg %d, old value %s, new value %s\n",
			dsc->tmp_regno, paddress (gdbarch, dsc->tmp_save),
			paddress (gdbarch, rip_base));
}

static void
fixup_displaced_copy (struct gdbarch *gdbarch,
		      amd64_displaced_step_closure *dsc,
		      CORE_ADDR from, CORE_ADDR to, struct regcache *regs)
{
  const struct amd64_insn *details = &dsc->insn_details;

  if (details->modrm_offset != -1)
    {
      gdb_byte modrm = details->raw_insn[details->modrm_offset];

      if ((modrm & 0xc7) == 0x05)
	{
	  /* The insn uses rip-relative addressing.
	     Deal with it.  */
	  fixup_riprel (gdbarch, dsc, from, to, regs);
	}
    }
}

struct displaced_step_closure *
amd64_displaced_step_copy_insn (struct gdbarch *gdbarch,
				CORE_ADDR from, CORE_ADDR to,
				struct regcache *regs)
{
  int len = gdbarch_max_insn_length (gdbarch);
  /* Extra space for sentinels so fixup_{riprel,displaced_copy} don't have to
     continually watch for running off the end of the buffer.  */
  int fixup_sentinel_space = len;
  amd64_displaced_step_closure *dsc
    = new amd64_displaced_step_closure (len + fixup_sentinel_space);
  gdb_byte *buf = &dsc->insn_buf[0];
  struct amd64_insn *details = &dsc->insn_details;

  read_memory (from, buf, len);

  /* Set up the sentinel space so we don't have to worry about running
     off the end of the buffer.  An excessive number of leading prefixes
     could otherwise cause this.  */
  memset (buf + len, 0, fixup_sentinel_space);

  amd64_get_insn_details (buf, details);

  /* GDB may get control back after the insn after the syscall.
     Presumably this is a kernel bug.
     If this is a syscall, make sure there's a nop afterwards.  */
  {
    int syscall_length;

    if (amd64_syscall_p (details, &syscall_length))
      buf[details->opcode_offset + syscall_length] = NOP_OPCODE;
  }

  /* Modify the insn to cope with the address where it will be executed from.
     In particular, handle any rip-relative addressing.	 */
  fixup_displaced_copy (gdbarch, dsc, from, to, regs);

  write_memory (to, buf, len);

  if (debug_displaced)
    {
      fprintf_unfiltered (gdb_stdlog, "displaced: copy %s->%s: ",
			  paddress (gdbarch, from), paddress (gdbarch, to));
      displaced_step_dump_bytes (gdb_stdlog, buf, len);
    }

  return dsc;
}

static int
amd64_absolute_jmp_p (const struct amd64_insn *details)
{
  const gdb_byte *insn = &details->raw_insn[details->opcode_offset];

  if (insn[0] == 0xff)
    {
      /* jump near, absolute indirect (/4) */
      if ((insn[1] & 0x38) == 0x20)
	return 1;

      /* jump far, absolute indirect (/5) */
      if ((insn[1] & 0x38) == 0x28)
	return 1;
    }

  return 0;
}

/* Return non-zero if the instruction DETAILS is a jump, zero otherwise.  */

static int
amd64_jmp_p (const struct amd64_insn *details)
{
  const gdb_byte *insn = &details->raw_insn[details->opcode_offset];

  /* jump short, relative.  */
  if (insn[0] == 0xeb)
    return 1;

  /* jump near, relative.  */
  if (insn[0] == 0xe9)
    return 1;

  return amd64_absolute_jmp_p (details);
}

static int
amd64_absolute_call_p (const struct amd64_insn *details)
{
  const gdb_byte *insn = &details->raw_insn[details->opcode_offset];

  if (insn[0] == 0xff)
    {
      /* Call near, absolute indirect (/2) */
      if ((insn[1] & 0x38) == 0x10)
	return 1;

      /* Call far, absolute indirect (/3) */
      if ((insn[1] & 0x38) == 0x18)
	return 1;
    }

  return 0;
}

static int
amd64_ret_p (const struct amd64_insn *details)
{
  /* NOTE: gcc can emit "repz ; ret".  */
  const gdb_byte *insn = &details->raw_insn[details->opcode_offset];

  switch (insn[0])
    {
    case 0xc2: /* ret near, pop N bytes */
    case 0xc3: /* ret near */
    case 0xca: /* ret far, pop N bytes */
    case 0xcb: /* ret far */
    case 0xcf: /* iret */
      return 1;

    default:
      return 0;
    }
}

static int
amd64_call_p (const struct amd64_insn *details)
{
  const gdb_byte *insn = &details->raw_insn[details->opcode_offset];

  if (amd64_absolute_call_p (details))
    return 1;

  /* call near, relative */
  if (insn[0] == 0xe8)
    return 1;

  return 0;
}

/* Return non-zero if INSN is a system call, and set *LENGTHP to its
   length in bytes.  Otherwise, return zero.  */

static int
amd64_syscall_p (const struct amd64_insn *details, int *lengthp)
{
  const gdb_byte *insn = &details->raw_insn[details->opcode_offset];

  if (insn[0] == 0x0f && insn[1] == 0x05)
    {
      *lengthp = 2;
      return 1;
    }

  return 0;
}

/* Classify the instruction at ADDR using PRED.
   Throw an error if the memory can't be read.  */

static int
amd64_classify_insn_at (struct gdbarch *gdbarch, CORE_ADDR addr,
			int (*pred) (const struct amd64_insn *))
{
  struct amd64_insn details;
  gdb_byte *buf;
  int len, classification;

  len = gdbarch_max_insn_length (gdbarch);
  buf = (gdb_byte *) alloca (len);

  read_code (addr, buf, len);
  amd64_get_insn_details (buf, &details);

  classification = pred (&details);

  return classification;
}

/* The gdbarch insn_is_call method.  */

static int
amd64_insn_is_call (struct gdbarch *gdbarch, CORE_ADDR addr)
{
  return amd64_classify_insn_at (gdbarch, addr, amd64_call_p);
}

/* The gdbarch insn_is_ret method.  */

static int
amd64_insn_is_ret (struct gdbarch *gdbarch, CORE_ADDR addr)
{
  return amd64_classify_insn_at (gdbarch, addr, amd64_ret_p);
}

/* The gdbarch insn_is_jump method.  */

static int
amd64_insn_is_jump (struct gdbarch *gdbarch, CORE_ADDR addr)
{
  return amd64_classify_insn_at (gdbarch, addr, amd64_jmp_p);
}

/* Fix up the state of registers and memory after having single-stepped
   a displaced instruction.  */

void
amd64_displaced_step_fixup (struct gdbarch *gdbarch,
			    struct displaced_step_closure *dsc_,
			    CORE_ADDR from, CORE_ADDR to,
			    struct regcache *regs)
{
  amd64_displaced_step_closure *dsc = (amd64_displaced_step_closure *) dsc_;
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  /* The offset we applied to the instruction's address.  */
  ULONGEST insn_offset = to - from;
  gdb_byte *insn = dsc->insn_buf.data ();
  const struct amd64_insn *insn_details = &dsc->insn_details;

  if (debug_displaced)
    fprintf_unfiltered (gdb_stdlog,
			"displaced: fixup (%s, %s), "
			"insn = 0x%02x 0x%02x ...\n",
			paddress (gdbarch, from), paddress (gdbarch, to),
			insn[0], insn[1]);

  /* If we used a tmp reg, restore it.	*/

  if (dsc->tmp_used)
    {
      if (debug_displaced)
	fprintf_unfiltered (gdb_stdlog, "displaced: restoring reg %d to %s\n",
			    dsc->tmp_regno, paddress (gdbarch, dsc->tmp_save));
      regcache_cooked_write_unsigned (regs, dsc->tmp_regno, dsc->tmp_save);
    }

  /* The list of issues to contend with here is taken from
     resume_execution in arch/x86/kernel/kprobes.c, Linux 2.6.28.
     Yay for Free Software!  */

  /* Relocate the %rip back to the program's instruction stream,
     if necessary.  */

  /* Except in the case of absolute or indirect jump or call
     instructions, or a return instruction, the new rip is relative to
     the displaced instruction; make it relative to the original insn.
     Well, signal handler returns don't need relocation either, but we use the
     value of %rip to recognize those; see below.  */
  if (! amd64_absolute_jmp_p (insn_details)
      && ! amd64_absolute_call_p (insn_details)
      && ! amd64_ret_p (insn_details))
    {
      ULONGEST orig_rip;
      int insn_len;

      regcache_cooked_read_unsigned (regs, AMD64_RIP_REGNUM, &orig_rip);

      /* A signal trampoline system call changes the %rip, resuming
	 execution of the main program after the signal handler has
	 returned.  That makes them like 'return' instructions; we
	 shouldn't relocate %rip.

	 But most system calls don't, and we do need to relocate %rip.

	 Our heuristic for distinguishing these cases: if stepping
	 over the system call instruction left control directly after
	 the instruction, the we relocate --- control almost certainly
	 doesn't belong in the displaced copy.	Otherwise, we assume
	 the instruction has put control where it belongs, and leave
	 it unrelocated.  Goodness help us if there are PC-relative
	 system calls.	*/
      if (amd64_syscall_p (insn_details, &insn_len)
	  && orig_rip != to + insn_len
	  /* GDB can get control back after the insn after the syscall.
	     Presumably this is a kernel bug.
	     Fixup ensures its a nop, we add one to the length for it.  */
	  && orig_rip != to + insn_len + 1)
	{
	  if (debug_displaced)
	    fprintf_unfiltered (gdb_stdlog,
				"displaced: syscall changed %%rip; "
				"not relocating\n");
	}
      else
	{
	  ULONGEST rip = orig_rip - insn_offset;

	  /* If we just stepped over a breakpoint insn, we don't backup
	     the pc on purpose; this is to match behaviour without
	     stepping.  */

	  regcache_cooked_write_unsigned (regs, AMD64_RIP_REGNUM, rip);

	  if (debug_displaced)
	    fprintf_unfiltered (gdb_stdlog,
				"displaced: "
				"relocated %%rip from %s to %s\n",
				paddress (gdbarch, orig_rip),
				paddress (gdbarch, rip));
	}
    }

  /* If the instruction was PUSHFL, then the TF bit will be set in the
     pushed value, and should be cleared.  We'll leave this for later,
     since GDB already messes up the TF flag when stepping over a
     pushfl.  */

  /* If the instruction was a call, the return address now atop the
     stack is the address following the copied instruction.  We need
     to make it the address following the original instruction.	 */
  if (amd64_call_p (insn_details))
    {
      ULONGEST rsp;
      ULONGEST retaddr;
      const ULONGEST retaddr_len = 8;

      regcache_cooked_read_unsigned (regs, AMD64_RSP_REGNUM, &rsp);
      retaddr = read_memory_unsigned_integer (rsp, retaddr_len, byte_order);
      retaddr = (retaddr - insn_offset) & 0xffffffffffffffffULL;
      write_memory_unsigned_integer (rsp, retaddr_len, byte_order, retaddr);

      if (debug_displaced)
	fprintf_unfiltered (gdb_stdlog,
			    "displaced: relocated return addr at %s "
			    "to %s\n",
			    paddress (gdbarch, rsp),
			    paddress (gdbarch, retaddr));
    }
}

/* If the instruction INSN uses RIP-relative addressing, return the
   offset into the raw INSN where the displacement to be adjusted is
   found.  Returns 0 if the instruction doesn't use RIP-relative
   addressing.  */

static int
rip_relative_offset (struct amd64_insn *insn)
{
  if (insn->modrm_offset != -1)
    {
      gdb_byte modrm = insn->raw_insn[insn->modrm_offset];

      if ((modrm & 0xc7) == 0x05)
	{
	  /* The displacement is found right after the ModRM byte.  */
	  return insn->modrm_offset + 1;
	}
    }

  return 0;
}

static void
append_insns (CORE_ADDR *to, ULONGEST len, const gdb_byte *buf)
{
  target_write_memory (*to, buf, len);
  *to += len;
}

static void
amd64_relocate_instruction (struct gdbarch *gdbarch,
			    CORE_ADDR *to, CORE_ADDR oldloc)
{
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  int len = gdbarch_max_insn_length (gdbarch);
  /* Extra space for sentinels.  */
  int fixup_sentinel_space = len;
  gdb_byte *buf = (gdb_byte *) xmalloc (len + fixup_sentinel_space);
  struct amd64_insn insn_details;
  int offset = 0;
  LONGEST rel32, newrel;
  gdb_byte *insn;
  int insn_length;

  read_memory (oldloc, buf, len);

  /* Set up the sentinel space so we don't have to worry about running
     off the end of the buffer.  An excessive number of leading prefixes
     could otherwise cause this.  */
  memset (buf + len, 0, fixup_sentinel_space);

  insn = buf;
  amd64_get_insn_details (insn, &insn_details);

  insn_length = gdb_buffered_insn_length (gdbarch, insn, len, oldloc);

  /* Skip legacy instruction prefixes.  */
  insn = amd64_skip_prefixes (insn);

  /* Adjust calls with 32-bit relative addresses as push/jump, with
     the address pushed being the location where the original call in
     the user program would return to.  */
  if (insn[0] == 0xe8)
    {
      gdb_byte push_buf[32];
      CORE_ADDR ret_addr;
      int i = 0;

      /* Where "ret" in the original code will return to.  */
      ret_addr = oldloc + insn_length;

      /* If pushing an address higher than or equal to 0x80000000,
	 avoid 'pushq', as that sign extends its 32-bit operand, which
	 would be incorrect.  */
      if (ret_addr <= 0x7fffffff)
	{
	  push_buf[0] = 0x68; /* pushq $...  */
	  store_unsigned_integer (&push_buf[1], 4, byte_order, ret_addr);
	  i = 5;
	}
      else
	{
	  push_buf[i++] = 0x48; /* sub    $0x8,%rsp */
	  push_buf[i++] = 0x83;
	  push_buf[i++] = 0xec;
	  push_buf[i++] = 0x08;

	  push_buf[i++] = 0xc7; /* movl    $imm,(%rsp) */
	  push_buf[i++] = 0x04;
	  push_buf[i++] = 0x24;
	  store_unsigned_integer (&push_buf[i], 4, byte_order,
				  ret_addr & 0xffffffff);
	  i += 4;

	  push_buf[i++] = 0xc7; /* movl    $imm,4(%rsp) */
	  push_buf[i++] = 0x44;
	  push_buf[i++] = 0x24;
	  push_buf[i++] = 0x04;
	  store_unsigned_integer (&push_buf[i], 4, byte_order,
				  ret_addr >> 32);
	  i += 4;
	}
      gdb_assert (i <= sizeof (push_buf));
      /* Push the push.  */
      append_insns (to, i, push_buf);

      /* Convert the relative call to a relative jump.  */
      insn[0] = 0xe9;

      /* Adjust the destination offset.  */
      rel32 = extract_signed_integer (insn + 1, 4, byte_order);
      newrel = (oldloc - *to) + rel32;
      store_signed_integer (insn + 1, 4, byte_order, newrel);

      if (debug_displaced)
	fprintf_unfiltered (gdb_stdlog,
			    "Adjusted insn rel32=%s at %s to"
			    " rel32=%s at %s\n",
			    hex_string (rel32), paddress (gdbarch, oldloc),
			    hex_string (newrel), paddress (gdbarch, *to));

      /* Write the adjusted jump into its displaced location.  */
      append_insns (to, 5, insn);
      return;
    }

  offset = rip_relative_offset (&insn_details);
  if (!offset)
    {
      /* Adjust jumps with 32-bit relative addresses.  Calls are
	 already handled above.  */
      if (insn[0] == 0xe9)
	offset = 1;
      /* Adjust conditional jumps.  */
      else if (insn[0] == 0x0f && (insn[1] & 0xf0) == 0x80)
	offset = 2;
    }

  if (offset)
    {
      rel32 = extract_signed_integer (insn + offset, 4, byte_order);
      newrel = (oldloc - *to) + rel32;
      store_signed_integer (insn + offset, 4, byte_order, newrel);
      if (debug_displaced)
	fprintf_unfiltered (gdb_stdlog,
			    "Adjusted insn rel32=%s at %s to"
			    " rel32=%s at %s\n",
			    hex_string (rel32), paddress (gdbarch, oldloc),
			    hex_string (newrel), paddress (gdbarch, *to));
    }

  /* Write the adjusted instruction into its displaced location.  */
  append_insns (to, insn_length, buf);
}


/* The maximum number of saved registers.  This should include %rip.  */
#define AMD64_NUM_SAVED_REGS	AMD64_NUM_GREGS

struct amd64_frame_cache
{
  /* Base address.  */
  CORE_ADDR base;
  int base_p;
  CORE_ADDR sp_offset;
  CORE_ADDR pc;

  /* Saved registers.  */
  CORE_ADDR saved_regs[AMD64_NUM_SAVED_REGS];
  CORE_ADDR saved_sp;
  int saved_sp_reg;

  /* Do we have a frame?  */
  int frameless_p;
};

/* Initialize a frame cache.  */

static void
amd64_init_frame_cache (struct amd64_frame_cache *cache)
{
  int i;

  /* Base address.  */
  cache->base = 0;
  cache->base_p = 0;
  cache->sp_offset = -8;
  cache->pc = 0;

  /* Saved registers.  We initialize these to -1 since zero is a valid
     offset (that's where %rbp is supposed to be stored).
     The values start out as being offsets, and are later converted to
     addresses (at which point -1 is interpreted as an address, still meaning
     "invalid").  */
  for (i = 0; i < AMD64_NUM_SAVED_REGS; i++)
    cache->saved_regs[i] = -1;
  cache->saved_sp = 0;
  cache->saved_sp_reg = -1;

  /* Frameless until proven otherwise.  */
  cache->frameless_p = 1;
}

/* Allocate and initialize a frame cache.  */

static struct amd64_frame_cache *
amd64_alloc_frame_cache (void)
{
  struct amd64_frame_cache *cache;

  cache = FRAME_OBSTACK_ZALLOC (struct amd64_frame_cache);
  amd64_init_frame_cache (cache);
  return cache;
}

/* GCC 4.4 and later, can put code in the prologue to realign the
   stack pointer.  Check whether PC points to such code, and update
   CACHE accordingly.  Return the first instruction after the code
   sequence or CURRENT_PC, whichever is smaller.  If we don't
   recognize the code, return PC.  */

static CORE_ADDR
amd64_analyze_stack_align (CORE_ADDR pc, CORE_ADDR current_pc,
			   struct amd64_frame_cache *cache)
{
  /* There are 2 code sequences to re-align stack before the frame
     gets set up:

	1. Use a caller-saved saved register:

		leaq  8(%rsp), %reg
		andq  $-XXX, %rsp
		pushq -8(%reg)

	2. Use a callee-saved saved register:

		pushq %reg
		leaq  16(%rsp), %reg
		andq  $-XXX, %rsp
		pushq -8(%reg)

     "andq $-XXX, %rsp" can be either 4 bytes or 7 bytes:
     
     	0x48 0x83 0xe4 0xf0			andq $-16, %rsp
     	0x48 0x81 0xe4 0x00 0xff 0xff 0xff	andq $-256, %rsp
   */

  gdb_byte buf[18];
  int reg, r;
  int offset, offset_and;

  if (target_read_code (pc, buf, sizeof buf))
    return pc;

  /* Check caller-saved saved register.  The first instruction has
     to be "leaq 8(%rsp), %reg".  */
  if ((buf[0] & 0xfb) == 0x48
      && buf[1] == 0x8d
      && buf[3] == 0x24
      && buf[4] == 0x8)
    {
      /* MOD must be binary 10 and R/M must be binary 100.  */
      if ((buf[2] & 0xc7) != 0x44)
	return pc;

      /* REG has register number.  */
      reg = (buf[2] >> 3) & 7;

      /* Check the REX.R bit.  */
      if (buf[0] == 0x4c)
	reg += 8;

      offset = 5;
    }
  else
    {
      /* Check callee-saved saved register.  The first instruction
	 has to be "pushq %reg".  */
      reg = 0;
      if ((buf[0] & 0xf8) == 0x50)
	offset = 0;
      else if ((buf[0] & 0xf6) == 0x40
	       && (buf[1] & 0xf8) == 0x50)
	{
	  /* Check the REX.B bit.  */
	  if ((buf[0] & 1) != 0)
	    reg = 8;

	  offset = 1;
	}
      else
	return pc;

      /* Get register.  */
      reg += buf[offset] & 0x7;

      offset++;

      /* The next instruction has to be "leaq 16(%rsp), %reg".  */
      if ((buf[offset] & 0xfb) != 0x48
	  || buf[offset + 1] != 0x8d
	  || buf[offset + 3] != 0x24
	  || buf[offset + 4] != 0x10)
	return pc;

      /* MOD must be binary 10 and R/M must be binary 100.  */
      if ((buf[offset + 2] & 0xc7) != 0x44)
	return pc;
      
      /* REG has register number.  */
      r = (buf[offset + 2] >> 3) & 7;

      /* Check the REX.R bit.  */
      if (buf[offset] == 0x4c)
	r += 8;

      /* Registers in pushq and leaq have to be the same.  */
      if (reg != r)
	return pc;

      offset += 5;
    }

  /* Rigister can't be %rsp nor %rbp.  */
  if (reg == 4 || reg == 5)
    return pc;

  /* The next instruction has to be "andq $-XXX, %rsp".  */
  if (buf[offset] != 0x48
      || buf[offset + 2] != 0xe4
      || (buf[offset + 1] != 0x81 && buf[offset + 1] != 0x83))
    return pc;

  offset_and = offset;
  offset += buf[offset + 1] == 0x81 ? 7 : 4;

  /* The next instruction has to be "pushq -8(%reg)".  */
  r = 0;
  if (buf[offset] == 0xff)
    offset++;
  else if ((buf[offset] & 0xf6) == 0x40
	   && buf[offset + 1] == 0xff)
    {
      /* Check the REX.B bit.  */
      if ((buf[offset] & 0x1) != 0)
	r = 8;
      offset += 2;
    }
  else
    return pc;

  /* 8bit -8 is 0xf8.  REG must be binary 110 and MOD must be binary
     01.  */
  if (buf[offset + 1] != 0xf8
      || (buf[offset] & 0xf8) != 0x70)
    return pc;

  /* R/M has register.  */
  r += buf[offset] & 7;

  /* Registers in leaq and pushq have to be the same.  */
  if (reg != r)
    return pc;

  if (current_pc > pc + offset_and)
    cache->saved_sp_reg = amd64_arch_reg_to_regnum (reg);

  return std::min (pc + offset + 2, current_pc);
}

/* Similar to amd64_analyze_stack_align for x32.  */

static CORE_ADDR
amd64_x32_analyze_stack_align (CORE_ADDR pc, CORE_ADDR current_pc,
			       struct amd64_frame_cache *cache) 
{
  /* There are 2 code sequences to re-align stack before the frame
     gets set up:

	1. Use a caller-saved saved register:

		leaq  8(%rsp), %reg
		andq  $-XXX, %rsp
		pushq -8(%reg)

	   or

		[addr32] leal  8(%rsp), %reg
		andl  $-XXX, %esp
		[addr32] pushq -8(%reg)

	2. Use a callee-saved saved register:

		pushq %reg
		leaq  16(%rsp), %reg
		andq  $-XXX, %rsp
		pushq -8(%reg)

	   or

		pushq %reg
		[addr32] leal  16(%rsp), %reg
		andl  $-XXX, %esp
		[addr32] pushq -8(%reg)

     "andq $-XXX, %rsp" can be either 4 bytes or 7 bytes:
     
     	0x48 0x83 0xe4 0xf0			andq $-16, %rsp
     	0x48 0x81 0xe4 0x00 0xff 0xff 0xff	andq $-256, %rsp

     "andl $-XXX, %esp" can be either 3 bytes or 6 bytes:
     
     	0x83 0xe4 0xf0			andl $-16, %esp
     	0x81 0xe4 0x00 0xff 0xff 0xff	andl $-256, %esp
   */

  gdb_byte buf[19];
  int reg, r;
  int offset, offset_and;

  if (target_read_memory (pc, buf, sizeof buf))
    return pc;

  /* Skip optional addr32 prefix.  */
  offset = buf[0] == 0x67 ? 1 : 0;

  /* Check caller-saved saved register.  The first instruction has
     to be "leaq 8(%rsp), %reg" or "leal 8(%rsp), %reg".  */
  if (((buf[offset] & 0xfb) == 0x48 || (buf[offset] & 0xfb) == 0x40)
      && buf[offset + 1] == 0x8d
      && buf[offset + 3] == 0x24
      && buf[offset + 4] == 0x8)
    {
      /* MOD must be binary 10 and R/M must be binary 100.  */
      if ((buf[offset + 2] & 0xc7) != 0x44)
	return pc;

      /* REG has register number.  */
      reg = (buf[offset + 2] >> 3) & 7;

      /* Check the REX.R bit.  */
      if ((buf[offset] & 0x4) != 0)
	reg += 8;

      offset += 5;
    }
  else
    {
      /* Check callee-saved saved register.  The first instruction
	 has to be "pushq %reg".  */
      reg = 0;
      if ((buf[offset] & 0xf6) == 0x40
	  && (buf[offset + 1] & 0xf8) == 0x50)
	{
	  /* Check the REX.B bit.  */
	  if ((buf[offset] & 1) != 0)
	    reg = 8;

	  offset += 1;
	}
      else if ((buf[offset] & 0xf8) != 0x50)
	return pc;

      /* Get register.  */
      reg += buf[offset] & 0x7;

      offset++;

      /* Skip optional addr32 prefix.  */
      if (buf[offset] == 0x67)
	offset++;

      /* The next instruction has to be "leaq 16(%rsp), %reg" or
	 "leal 16(%rsp), %reg".  */
      if (((buf[offset] & 0xfb) != 0x48 && (buf[offset] & 0xfb) != 0x40)
	  || buf[offset + 1] != 0x8d
	  || buf[offset + 3] != 0x24
	  || buf[offset + 4] != 0x10)
	return pc;

      /* MOD must be binary 10 and R/M must be binary 100.  */
      if ((buf[offset + 2] & 0xc7) != 0x44)
	return pc;
      
      /* REG has register number.  */
      r = (buf[offset + 2] >> 3) & 7;

      /* Check the REX.R bit.  */
      if ((buf[offset] & 0x4) != 0)
	r += 8;

      /* Registers in pushq and leaq have to be the same.  */
      if (reg != r)
	return pc;

      offset += 5;
    }

  /* Rigister can't be %rsp nor %rbp.  */
  if (reg == 4 || reg == 5)
    return pc;

  /* The next instruction may be "andq $-XXX, %rsp" or
     "andl $-XXX, %esp".  */
  if (buf[offset] != 0x48)
    offset--;

  if (buf[offset + 2] != 0xe4
      || (buf[offset + 1] != 0x81 && buf[offset + 1] != 0x83))
    return pc;

  offset_and = offset;
  offset += buf[offset + 1] == 0x81 ? 7 : 4;

  /* Skip optional addr32 prefix.  */
  if (buf[offset] == 0x67)
    offset++;

  /* The next instruction has to be "pushq -8(%reg)".  */
  r = 0;
  if (buf[offset] == 0xff)
    offset++;
  else if ((buf[offset] & 0xf6) == 0x40
	   && buf[offset + 1] == 0xff)
    {
      /* Check the REX.B bit.  */
      if ((buf[offset] & 0x1) != 0)
	r = 8;
      offset += 2;
    }
  else
    return pc;

  /* 8bit -8 is 0xf8.  REG must be binary 110 and MOD must be binary
     01.  */
  if (buf[offset + 1] != 0xf8
      || (buf[offset] & 0xf8) != 0x70)
    return pc;

  /* R/M has register.  */
  r += buf[offset] & 7;

  /* Registers in leaq and pushq have to be the same.  */
  if (reg != r)
    return pc;

  if (current_pc > pc + offset_and)
    cache->saved_sp_reg = amd64_arch_reg_to_regnum (reg);

  return std::min (pc + offset + 2, current_pc);
}

/* Do a limited analysis of the prologue at PC and update CACHE
   accordingly.  Bail out early if CURRENT_PC is reached.  Return the
   address where the analysis stopped.

   We will handle only functions beginning with:

      pushq %rbp        0x55
      movq %rsp, %rbp   0x48 0x89 0xe5 (or 0x48 0x8b 0xec)

   or (for the X32 ABI):

      pushq %rbp        0x55
      movl %esp, %ebp   0x89 0xe5 (or 0x8b 0xec)

   Any function that doesn't start with one of these sequences will be
   assumed to have no prologue and thus no valid frame pointer in
   %rbp.  */

static CORE_ADDR
amd64_analyze_prologue (struct gdbarch *gdbarch,
			CORE_ADDR pc, CORE_ADDR current_pc,
			struct amd64_frame_cache *cache)
{
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  /* There are two variations of movq %rsp, %rbp.  */
  static const gdb_byte mov_rsp_rbp_1[3] = { 0x48, 0x89, 0xe5 };
  static const gdb_byte mov_rsp_rbp_2[3] = { 0x48, 0x8b, 0xec };
  /* Ditto for movl %esp, %ebp.  */
  static const gdb_byte mov_esp_ebp_1[2] = { 0x89, 0xe5 };
  static const gdb_byte mov_esp_ebp_2[2] = { 0x8b, 0xec };

  gdb_byte buf[3];
  gdb_byte op;

  if (current_pc <= pc)
    return current_pc;

  if (gdbarch_ptr_bit (gdbarch) == 32)
    pc = amd64_x32_analyze_stack_align (pc, current_pc, cache);
  else
    pc = amd64_analyze_stack_align (pc, current_pc, cache);

  op = read_code_unsigned_integer (pc, 1, byte_order);

  if (op == 0x55)		/* pushq %rbp */
    {
      /* Take into account that we've executed the `pushq %rbp' that
         starts this instruction sequence.  */
      cache->saved_regs[AMD64_RBP_REGNUM] = 0;
      cache->sp_offset += 8;

      /* If that's all, return now.  */
      if (current_pc <= pc + 1)
        return current_pc;

      read_code (pc + 1, buf, 3);

      /* Check for `movq %rsp, %rbp'.  */
      if (memcmp (buf, mov_rsp_rbp_1, 3) == 0
	  || memcmp (buf, mov_rsp_rbp_2, 3) == 0)
	{
	  /* OK, we actually have a frame.  */
	  cache->frameless_p = 0;
	  return pc + 4;
	}

      /* For X32, also check for `movq %esp, %ebp'.  */
      if (gdbarch_ptr_bit (gdbarch) == 32)
	{
	  if (memcmp (buf, mov_esp_ebp_1, 2) == 0
	      || memcmp (buf, mov_esp_ebp_2, 2) == 0)
	    {
	      /* OK, we actually have a frame.  */
	      cache->frameless_p = 0;
	      return pc + 3;
	    }
	}

      return pc + 1;
    }

  return pc;
}

/* Work around false termination of prologue - GCC PR debug/48827.

   START_PC is the first instruction of a function, PC is its minimal already
   determined advanced address.  Function returns PC if it has nothing to do.

   84 c0                test   %al,%al
   74 23                je     after
   <-- here is 0 lines advance - the false prologue end marker.
   0f 29 85 70 ff ff ff movaps %xmm0,-0x90(%rbp)
   0f 29 4d 80          movaps %xmm1,-0x80(%rbp)
   0f 29 55 90          movaps %xmm2,-0x70(%rbp)
   0f 29 5d a0          movaps %xmm3,-0x60(%rbp)
   0f 29 65 b0          movaps %xmm4,-0x50(%rbp)
   0f 29 6d c0          movaps %xmm5,-0x40(%rbp)
   0f 29 75 d0          movaps %xmm6,-0x30(%rbp)
   0f 29 7d e0          movaps %xmm7,-0x20(%rbp)
   after:  */

static CORE_ADDR
amd64_skip_xmm_prologue (CORE_ADDR pc, CORE_ADDR start_pc)
{
  struct symtab_and_line start_pc_sal, next_sal;
  gdb_byte buf[4 + 8 * 7];
  int offset, xmmreg;

  if (pc == start_pc)
    return pc;

  start_pc_sal = find_pc_sect_line (start_pc, NULL, 0);
  if (start_pc_sal.symtab == NULL
      || producer_is_gcc_ge_4 (COMPUNIT_PRODUCER
	   (SYMTAB_COMPUNIT (start_pc_sal.symtab))) < 6
      || start_pc_sal.pc != start_pc || pc >= start_pc_sal.end)
    return pc;

  next_sal = find_pc_sect_line (start_pc_sal.end, NULL, 0);
  if (next_sal.line != start_pc_sal.line)
    return pc;

  /* START_PC can be from overlayed memory, ignored here.  */
  if (target_read_code (next_sal.pc - 4, buf, sizeof (buf)) != 0)
    return pc;

  /* test %al,%al */
  if (buf[0] != 0x84 || buf[1] != 0xc0)
    return pc;
  /* je AFTER */
  if (buf[2] != 0x74)
    return pc;

  offset = 4;
  for (xmmreg = 0; xmmreg < 8; xmmreg++)
    {
      /* 0x0f 0x29 0b??000101 movaps %xmmreg?,-0x??(%rbp) */
      if (buf[offset] != 0x0f || buf[offset + 1] != 0x29
          || (buf[offset + 2] & 0x3f) != (xmmreg << 3 | 0x5))
	return pc;

      /* 0b01?????? */
      if ((buf[offset + 2] & 0xc0) == 0x40)
	{
	  /* 8-bit displacement.  */
	  offset += 4;
	}
      /* 0b10?????? */
      else if ((buf[offset + 2] & 0xc0) == 0x80)
	{
	  /* 32-bit displacement.  */
	  offset += 7;
	}
      else
	return pc;
    }

  /* je AFTER */
  if (offset - 4 != buf[3])
    return pc;

  return next_sal.end;
}

/* Return PC of first real instruction.  */

static CORE_ADDR
amd64_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
{
  struct amd64_frame_cache cache;
  CORE_ADDR pc;
  CORE_ADDR func_addr;

  if (find_pc_partial_function (start_pc, NULL, &func_addr, NULL))
    {
      CORE_ADDR post_prologue_pc
	= skip_prologue_using_sal (gdbarch, func_addr);
      struct compunit_symtab *cust = find_pc_compunit_symtab (func_addr);

      /* Clang always emits a line note before the prologue and another
	 one after.  We trust clang to emit usable line notes.  */
      if (post_prologue_pc
	  && (cust != NULL
	      && COMPUNIT_PRODUCER (cust) != NULL
	      && startswith (COMPUNIT_PRODUCER (cust), "clang ")))
        return std::max (start_pc, post_prologue_pc);
    }

  amd64_init_frame_cache (&cache);
  pc = amd64_analyze_prologue (gdbarch, start_pc, 0xffffffffffffffffLL,
			       &cache);
  if (cache.frameless_p)
    return start_pc;

  return amd64_skip_xmm_prologue (pc, start_pc);
}


/* Normal frames.  */

static void
amd64_frame_cache_1 (struct frame_info *this_frame,
		     struct amd64_frame_cache *cache)
{
  struct gdbarch *gdbarch = get_frame_arch (this_frame);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  gdb_byte buf[8];
  int i;

  cache->pc = get_frame_func (this_frame);
  if (cache->pc != 0)
    amd64_analyze_prologue (gdbarch, cache->pc, get_frame_pc (this_frame),
			    cache);

  if (cache->frameless_p)
    {
      /* We didn't find a valid frame.  If we're at the start of a
	 function, or somewhere half-way its prologue, the function's
	 frame probably hasn't been fully setup yet.  Try to
	 reconstruct the base address for the stack frame by looking
	 at the stack pointer.  For truly "frameless" functions this
	 might work too.  */

      if (cache->saved_sp_reg != -1)
	{
	  /* Stack pointer has been saved.  */
	  get_frame_register (this_frame, cache->saved_sp_reg, buf);
	  cache->saved_sp = extract_unsigned_integer (buf, 8, byte_order);

	  /* We're halfway aligning the stack.  */
	  cache->base = ((cache->saved_sp - 8) & 0xfffffffffffffff0LL) - 8;
	  cache->saved_regs[AMD64_RIP_REGNUM] = cache->saved_sp - 8;

	  /* This will be added back below.  */
	  cache->saved_regs[AMD64_RIP_REGNUM] -= cache->base;
	}
      else
	{
	  get_frame_register (this_frame, AMD64_RSP_REGNUM, buf);
	  cache->base = extract_unsigned_integer (buf, 8, byte_order)
			+ cache->sp_offset;
	}
    }
  else
    {
      get_frame_register (this_frame, AMD64_RBP_REGNUM, buf);
      cache->base = extract_unsigned_integer (buf, 8, byte_order);
    }

  /* Now that we have the base address for the stack frame we can
     calculate the value of %rsp in the calling frame.  */
  cache->saved_sp = cache->base + 16;

  /* For normal frames, %rip is stored at 8(%rbp).  If we don't have a
     frame we find it at the same offset from the reconstructed base
     address.  If we're halfway aligning the stack, %rip is handled
     differently (see above).  */
  if (!cache->frameless_p || cache->saved_sp_reg == -1)
    cache->saved_regs[AMD64_RIP_REGNUM] = 8;

  /* Adjust all the saved registers such that they contain addresses
     instead of offsets.  */
  for (i = 0; i < AMD64_NUM_SAVED_REGS; i++)
    if (cache->saved_regs[i] != -1)
      cache->saved_regs[i] += cache->base;

  cache->base_p = 1;
}

static struct amd64_frame_cache *
amd64_frame_cache (struct frame_info *this_frame, void **this_cache)
{
  struct amd64_frame_cache *cache;

  if (*this_cache)
    return (struct amd64_frame_cache *) *this_cache;

  cache = amd64_alloc_frame_cache ();
  *this_cache = cache;

  TRY
    {
      amd64_frame_cache_1 (this_frame, cache);
    }
  CATCH (ex, RETURN_MASK_ERROR)
    {
      if (ex.error != NOT_AVAILABLE_ERROR)
	throw_exception (ex);
    }
  END_CATCH

  return cache;
}

static enum unwind_stop_reason
amd64_frame_unwind_stop_reason (struct frame_info *this_frame,
				void **this_cache)
{
  struct amd64_frame_cache *cache =
    amd64_frame_cache (this_frame, this_cache);

  if (!cache->base_p)
    return UNWIND_UNAVAILABLE;

  /* This marks the outermost frame.  */
  if (cache->base == 0)
    return UNWIND_OUTERMOST;

  return UNWIND_NO_REASON;
}

static void
amd64_frame_this_id (struct frame_info *this_frame, void **this_cache,
		     struct frame_id *this_id)
{
  struct amd64_frame_cache *cache =
    amd64_frame_cache (this_frame, this_cache);

  if (!cache->base_p)
    (*this_id) = frame_id_build_unavailable_stack (cache->pc);
  else if (cache->base == 0)
    {
      /* This marks the outermost frame.  */
      return;
    }
  else
    (*this_id) = frame_id_build (cache->base + 16, cache->pc);
}

static struct value *
amd64_frame_prev_register (struct frame_info *this_frame, void **this_cache,
			   int regnum)
{
  struct gdbarch *gdbarch = get_frame_arch (this_frame);
  struct amd64_frame_cache *cache =
    amd64_frame_cache (this_frame, this_cache);

  gdb_assert (regnum >= 0);

  if (regnum == gdbarch_sp_regnum (gdbarch) && cache->saved_sp)
    return frame_unwind_got_constant (this_frame, regnum, cache->saved_sp);

  if (regnum < AMD64_NUM_SAVED_REGS && cache->saved_regs[regnum] != -1)
    return frame_unwind_got_memory (this_frame, regnum,
				    cache->saved_regs[regnum]);

  return frame_unwind_got_register (this_frame, regnum, regnum);
}

static const struct frame_unwind amd64_frame_unwind =
{
  NORMAL_FRAME,
  amd64_frame_unwind_stop_reason,
  amd64_frame_this_id,
  amd64_frame_prev_register,
  NULL,
  default_frame_sniffer
};

/* Generate a bytecode expression to get the value of the saved PC.  */

static void
amd64_gen_return_address (struct gdbarch *gdbarch,
			  struct agent_expr *ax, struct axs_value *value,
			  CORE_ADDR scope)
{
  /* The following sequence assumes the traditional use of the base
     register.  */
  ax_reg (ax, AMD64_RBP_REGNUM);
  ax_const_l (ax, 8);
  ax_simple (ax, aop_add);
  value->type = register_type (gdbarch, AMD64_RIP_REGNUM);
  value->kind = axs_lvalue_memory;
}


/* Signal trampolines.  */

/* FIXME: kettenis/20030419: Perhaps, we can unify the 32-bit and
   64-bit variants.  This would require using identical frame caches
   on both platforms.  */

static struct amd64_frame_cache *
amd64_sigtramp_frame_cache (struct frame_info *this_frame, void **this_cache)
{
  struct gdbarch *gdbarch = get_frame_arch (this_frame);
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  struct amd64_frame_cache *cache;
  CORE_ADDR addr;
  gdb_byte buf[8];
  int i;

  if (*this_cache)
    return (struct amd64_frame_cache *) *this_cache;

  cache = amd64_alloc_frame_cache ();

  TRY
    {
      get_frame_register (this_frame, AMD64_RSP_REGNUM, buf);
      cache->base = extract_unsigned_integer (buf, 8, byte_order) - 8;

      addr = tdep->sigcontext_addr (this_frame);
      gdb_assert (tdep->sc_reg_offset);
      gdb_assert (tdep->sc_num_regs <= AMD64_NUM_SAVED_REGS);
      for (i = 0; i < tdep->sc_num_regs; i++)
	if (tdep->sc_reg_offset[i] != -1)
	  cache->saved_regs[i] = addr + tdep->sc_reg_offset[i];

      cache->base_p = 1;
    }
  CATCH (ex, RETURN_MASK_ERROR)
    {
      if (ex.error != NOT_AVAILABLE_ERROR)
	throw_exception (ex);
    }
  END_CATCH

  *this_cache = cache;
  return cache;
}

static enum unwind_stop_reason
amd64_sigtramp_frame_unwind_stop_reason (struct frame_info *this_frame,
					 void **this_cache)
{
  struct amd64_frame_cache *cache =
    amd64_sigtramp_frame_cache (this_frame, this_cache);

  if (!cache->base_p)
    return UNWIND_UNAVAILABLE;

  return UNWIND_NO_REASON;
}

static void
amd64_sigtramp_frame_this_id (struct frame_info *this_frame,
			      void **this_cache, struct frame_id *this_id)
{
  struct amd64_frame_cache *cache =
    amd64_sigtramp_frame_cache (this_frame, this_cache);

  if (!cache->base_p)
    (*this_id) = frame_id_build_unavailable_stack (get_frame_pc (this_frame));
  else if (cache->base == 0)
    {
      /* This marks the outermost frame.  */
      return;
    }
  else
    (*this_id) = frame_id_build (cache->base + 16, get_frame_pc (this_frame));
}

static struct value *
amd64_sigtramp_frame_prev_register (struct frame_info *this_frame,
				    void **this_cache, int regnum)
{
  /* Make sure we've initialized the cache.  */
  amd64_sigtramp_frame_cache (this_frame, this_cache);

  return amd64_frame_prev_register (this_frame, this_cache, regnum);
}

static int
amd64_sigtramp_frame_sniffer (const struct frame_unwind *self,
			      struct frame_info *this_frame,
			      void **this_cache)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (get_frame_arch (this_frame));

  /* We shouldn't even bother if we don't have a sigcontext_addr
     handler.  */
  if (tdep->sigcontext_addr == NULL)
    return 0;

  if (tdep->sigtramp_p != NULL)
    {
      if (tdep->sigtramp_p (this_frame))
	return 1;
    }

  if (tdep->sigtramp_start != 0)
    {
      CORE_ADDR pc = get_frame_pc (this_frame);

      gdb_assert (tdep->sigtramp_end != 0);
      if (pc >= tdep->sigtramp_start && pc < tdep->sigtramp_end)
	return 1;
    }

  return 0;
}

static const struct frame_unwind amd64_sigtramp_frame_unwind =
{
  SIGTRAMP_FRAME,
  amd64_sigtramp_frame_unwind_stop_reason,
  amd64_sigtramp_frame_this_id,
  amd64_sigtramp_frame_prev_register,
  NULL,
  amd64_sigtramp_frame_sniffer
};


static CORE_ADDR
amd64_frame_base_address (struct frame_info *this_frame, void **this_cache)
{
  struct amd64_frame_cache *cache =
    amd64_frame_cache (this_frame, this_cache);

  return cache->base;
}

static const struct frame_base amd64_frame_base =
{
  &amd64_frame_unwind,
  amd64_frame_base_address,
  amd64_frame_base_address,
  amd64_frame_base_address
};

/* Normal frames, but in a function epilogue.  */

/* Implement the stack_frame_destroyed_p gdbarch method.

   The epilogue is defined here as the 'ret' instruction, which will
   follow any instruction such as 'leave' or 'pop %ebp' that destroys
   the function's stack frame.  */

static int
amd64_stack_frame_destroyed_p (struct gdbarch *gdbarch, CORE_ADDR pc)
{
  gdb_byte insn;
  struct compunit_symtab *cust;

  cust = find_pc_compunit_symtab (pc);
  if (cust != NULL && COMPUNIT_EPILOGUE_UNWIND_VALID (cust))
    return 0;

  if (target_read_memory (pc, &insn, 1))
    return 0;   /* Can't read memory at pc.  */

  if (insn != 0xc3)     /* 'ret' instruction.  */
    return 0;

  return 1;
}

static int
amd64_epilogue_frame_sniffer (const struct frame_unwind *self,
			      struct frame_info *this_frame,
			      void **this_prologue_cache)
{
  if (frame_relative_level (this_frame) == 0)
    return amd64_stack_frame_destroyed_p (get_frame_arch (this_frame),
					  get_frame_pc (this_frame));
  else
    return 0;
}

static struct amd64_frame_cache *
amd64_epilogue_frame_cache (struct frame_info *this_frame, void **this_cache)
{
  struct gdbarch *gdbarch = get_frame_arch (this_frame);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  struct amd64_frame_cache *cache;
  gdb_byte buf[8];

  if (*this_cache)
    return (struct amd64_frame_cache *) *this_cache;

  cache = amd64_alloc_frame_cache ();
  *this_cache = cache;

  TRY
    {
      /* Cache base will be %esp plus cache->sp_offset (-8).  */
      get_frame_register (this_frame, AMD64_RSP_REGNUM, buf);
      cache->base = extract_unsigned_integer (buf, 8,
					      byte_order) + cache->sp_offset;

      /* Cache pc will be the frame func.  */
      cache->pc = get_frame_pc (this_frame);

      /* The saved %esp will be at cache->base plus 16.  */
      cache->saved_sp = cache->base + 16;

      /* The saved %eip will be at cache->base plus 8.  */
      cache->saved_regs[AMD64_RIP_REGNUM] = cache->base + 8;

      cache->base_p = 1;
    }
  CATCH (ex, RETURN_MASK_ERROR)
    {
      if (ex.error != NOT_AVAILABLE_ERROR)
	throw_exception (ex);
    }
  END_CATCH

  return cache;
}

static enum unwind_stop_reason
amd64_epilogue_frame_unwind_stop_reason (struct frame_info *this_frame,
					 void **this_cache)
{
  struct amd64_frame_cache *cache
    = amd64_epilogue_frame_cache (this_frame, this_cache);

  if (!cache->base_p)
    return UNWIND_UNAVAILABLE;

  return UNWIND_NO_REASON;
}

static void
amd64_epilogue_frame_this_id (struct frame_info *this_frame,
			      void **this_cache,
			      struct frame_id *this_id)
{
  struct amd64_frame_cache *cache = amd64_epilogue_frame_cache (this_frame,
							       this_cache);

  if (!cache->base_p)
    (*this_id) = frame_id_build_unavailable_stack (cache->pc);
  else
    (*this_id) = frame_id_build (cache->base + 8, cache->pc);
}

static const struct frame_unwind amd64_epilogue_frame_unwind =
{
  NORMAL_FRAME,
  amd64_epilogue_frame_unwind_stop_reason,
  amd64_epilogue_frame_this_id,
  amd64_frame_prev_register,
  NULL, 
  amd64_epilogue_frame_sniffer
};

static struct frame_id
amd64_dummy_id (struct gdbarch *gdbarch, struct frame_info *this_frame)
{
  CORE_ADDR fp;

  fp = get_frame_register_unsigned (this_frame, AMD64_RBP_REGNUM);

  return frame_id_build (fp + 16, get_frame_pc (this_frame));
}

/* 16 byte align the SP per frame requirements.  */

static CORE_ADDR
amd64_frame_align (struct gdbarch *gdbarch, CORE_ADDR sp)
{
  return sp & -(CORE_ADDR)16;
}


/* Supply register REGNUM from the buffer specified by FPREGS and LEN
   in the floating-point register set REGSET to register cache
   REGCACHE.  If REGNUM is -1, do this for all registers in REGSET.  */

static void
amd64_supply_fpregset (const struct regset *regset, struct regcache *regcache,
		       int regnum, const void *fpregs, size_t len)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  const struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  gdb_assert (len >= tdep->sizeof_fpregset);
  amd64_supply_fxsave (regcache, regnum, fpregs);
}

/* Collect register REGNUM from the register cache REGCACHE and store
   it in the buffer specified by FPREGS and LEN as described by the
   floating-point register set REGSET.  If REGNUM is -1, do this for
   all registers in REGSET.  */

static void
amd64_collect_fpregset (const struct regset *regset,
			const struct regcache *regcache,
			int regnum, void *fpregs, size_t len)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  const struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  gdb_assert (len >= tdep->sizeof_fpregset);
  amd64_collect_fxsave (regcache, regnum, fpregs);
}

const struct regset amd64_fpregset =
  {
    NULL, amd64_supply_fpregset, amd64_collect_fpregset
  };


/* Figure out where the longjmp will land.  Slurp the jmp_buf out of
   %rdi.  We expect its value to be a pointer to the jmp_buf structure
   from which we extract the address that we will land at.  This
   address is copied into PC.  This routine returns non-zero on
   success.  */

static int
amd64_get_longjmp_target (struct frame_info *frame, CORE_ADDR *pc)
{
  gdb_byte buf[8];
  CORE_ADDR jb_addr;
  struct gdbarch *gdbarch = get_frame_arch (frame);
  int jb_pc_offset = gdbarch_tdep (gdbarch)->jb_pc_offset;
  int len = TYPE_LENGTH (builtin_type (gdbarch)->builtin_func_ptr);

  /* If JB_PC_OFFSET is -1, we have no way to find out where the
     longjmp will land.	 */
  if (jb_pc_offset == -1)
    return 0;

  get_frame_register (frame, AMD64_RDI_REGNUM, buf);
  jb_addr= extract_typed_address
	    (buf, builtin_type (gdbarch)->builtin_data_ptr);
  if (target_read_memory (jb_addr + jb_pc_offset, buf, len))
    return 0;

  *pc = extract_typed_address (buf, builtin_type (gdbarch)->builtin_func_ptr);

  return 1;
}

static const int amd64_record_regmap[] =
{
  AMD64_RAX_REGNUM, AMD64_RCX_REGNUM, AMD64_RDX_REGNUM, AMD64_RBX_REGNUM,
  AMD64_RSP_REGNUM, AMD64_RBP_REGNUM, AMD64_RSI_REGNUM, AMD64_RDI_REGNUM,
  AMD64_R8_REGNUM, AMD64_R9_REGNUM, AMD64_R10_REGNUM, AMD64_R11_REGNUM,
  AMD64_R12_REGNUM, AMD64_R13_REGNUM, AMD64_R14_REGNUM, AMD64_R15_REGNUM,
  AMD64_RIP_REGNUM, AMD64_EFLAGS_REGNUM, AMD64_CS_REGNUM, AMD64_SS_REGNUM,
  AMD64_DS_REGNUM, AMD64_ES_REGNUM, AMD64_FS_REGNUM, AMD64_GS_REGNUM
};

void
amd64_init_abi (struct gdbarch_info info, struct gdbarch *gdbarch,
		const target_desc *default_tdesc)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  const struct target_desc *tdesc = info.target_desc;
  static const char *const stap_integer_prefixes[] = { "$", NULL };
  static const char *const stap_register_prefixes[] = { "%", NULL };
  static const char *const stap_register_indirection_prefixes[] = { "(",
								    NULL };
  static const char *const stap_register_indirection_suffixes[] = { ")",
								    NULL };

  /* AMD64 generally uses `fxsave' instead of `fsave' for saving its
     floating-point registers.  */
  tdep->sizeof_fpregset = I387_SIZEOF_FXSAVE;
  tdep->fpregset = &amd64_fpregset;

  if (! tdesc_has_registers (tdesc))
    tdesc = default_tdesc;
  tdep->tdesc = tdesc;

  tdep->num_core_regs = AMD64_NUM_GREGS + I387_NUM_REGS;
  tdep->register_names = amd64_register_names;

  if (tdesc_find_feature (tdesc, "org.gnu.gdb.i386.avx512") != NULL)
    {
      tdep->zmmh_register_names = amd64_zmmh_names;
      tdep->k_register_names = amd64_k_names;
      tdep->xmm_avx512_register_names = amd64_xmm_avx512_names;
      tdep->ymm16h_register_names = amd64_ymmh_avx512_names;

      tdep->num_zmm_regs = 32;
      tdep->num_xmm_avx512_regs = 16;
      tdep->num_ymm_avx512_regs = 16;

      tdep->zmm0h_regnum = AMD64_ZMM0H_REGNUM;
      tdep->k0_regnum = AMD64_K0_REGNUM;
      tdep->xmm16_regnum = AMD64_XMM16_REGNUM;
      tdep->ymm16h_regnum = AMD64_YMM16H_REGNUM;
    }

  if (tdesc_find_feature (tdesc, "org.gnu.gdb.i386.avx") != NULL)
    {
      tdep->ymmh_register_names = amd64_ymmh_names;
      tdep->num_ymm_regs = 16;
      tdep->ymm0h_regnum = AMD64_YMM0H_REGNUM;
    }

  if (tdesc_find_feature (tdesc, "org.gnu.gdb.i386.mpx") != NULL)
    {
      tdep->mpx_register_names = amd64_mpx_names;
      tdep->bndcfgu_regnum = AMD64_BNDCFGU_REGNUM;
      tdep->bnd0r_regnum = AMD64_BND0R_REGNUM;
    }

  if (tdesc_find_feature (tdesc, "org.gnu.gdb.i386.segments") != NULL)
    {
      const struct tdesc_feature *feature =
	  tdesc_find_feature (tdesc, "org.gnu.gdb.i386.segments");
      struct tdesc_arch_data *tdesc_data_segments =
	  (struct tdesc_arch_data *) info.tdep_info;

      tdesc_numbered_register (feature, tdesc_data_segments,
		       AMD64_FSBASE_REGNUM, "fs_base");
      tdesc_numbered_register (feature, tdesc_data_segments,
		       AMD64_GSBASE_REGNUM, "gs_base");
    }

  if (tdesc_find_feature (tdesc, "org.gnu.gdb.i386.pkeys") != NULL)
    {
      tdep->pkeys_register_names = amd64_pkeys_names;
      tdep->pkru_regnum = AMD64_PKRU_REGNUM;
      tdep->num_pkeys_regs = 1;
    }

  tdep->num_byte_regs = 20;
  tdep->num_word_regs = 16;
  tdep->num_dword_regs = 16;
  /* Avoid wiring in the MMX registers for now.  */
  tdep->num_mmx_regs = 0;

  set_gdbarch_pseudo_register_read_value (gdbarch,
					  amd64_pseudo_register_read_value);
  set_gdbarch_pseudo_register_write (gdbarch,
				     amd64_pseudo_register_write);
  set_gdbarch_ax_pseudo_register_collect (gdbarch,
					  amd64_ax_pseudo_register_collect);

  set_tdesc_pseudo_register_name (gdbarch, amd64_pseudo_register_name);

  /* AMD64 has an FPU and 16 SSE registers.  */
  tdep->st0_regnum = AMD64_ST0_REGNUM;
  tdep->num_xmm_regs = 16;

  /* This is what all the fuss is about.  */
  set_gdbarch_long_bit (gdbarch, 64);
  set_gdbarch_long_long_bit (gdbarch, 64);
  set_gdbarch_ptr_bit (gdbarch, 64);

  /* In contrast to the i386, on AMD64 a `long double' actually takes
     up 128 bits, even though it's still based on the i387 extended
     floating-point format which has only 80 significant bits.  */
  set_gdbarch_long_double_bit (gdbarch, 128);

  set_gdbarch_num_regs (gdbarch, AMD64_NUM_REGS);

  /* Register numbers of various important registers.  */
  set_gdbarch_sp_regnum (gdbarch, AMD64_RSP_REGNUM); /* %rsp */
  set_gdbarch_pc_regnum (gdbarch, AMD64_RIP_REGNUM); /* %rip */
  set_gdbarch_ps_regnum (gdbarch, AMD64_EFLAGS_REGNUM); /* %eflags */
  set_gdbarch_fp0_regnum (gdbarch, AMD64_ST0_REGNUM); /* %st(0) */

  /* The "default" register numbering scheme for AMD64 is referred to
     as the "DWARF Register Number Mapping" in the System V psABI.
     The preferred debugging format for all known AMD64 targets is
     actually DWARF2, and GCC doesn't seem to support DWARF (that is
     DWARF-1), but we provide the same mapping just in case.  This
     mapping is also used for stabs, which GCC does support.  */
  set_gdbarch_stab_reg_to_regnum (gdbarch, amd64_dwarf_reg_to_regnum);
  set_gdbarch_dwarf2_reg_to_regnum (gdbarch, amd64_dwarf_reg_to_regnum);

  /* We don't override SDB_REG_RO_REGNUM, since COFF doesn't seem to
     be in use on any of the supported AMD64 targets.  */

  /* Call dummy code.  */
  set_gdbarch_push_dummy_call (gdbarch, amd64_push_dummy_call);
  set_gdbarch_frame_align (gdbarch, amd64_frame_align);
  set_gdbarch_frame_red_zone_size (gdbarch, 128);

  set_gdbarch_convert_register_p (gdbarch, i387_convert_register_p);
  set_gdbarch_register_to_value (gdbarch, i387_register_to_value);
  set_gdbarch_value_to_register (gdbarch, i387_value_to_register);

  set_gdbarch_return_value (gdbarch, amd64_return_value);

  set_gdbarch_skip_prologue (gdbarch, amd64_skip_prologue);

  tdep->record_regmap = amd64_record_regmap;

  set_gdbarch_dummy_id (gdbarch, amd64_dummy_id);

  /* Hook the function epilogue frame unwinder.  This unwinder is
     appended to the list first, so that it supercedes the other
     unwinders in function epilogues.  */
  frame_unwind_prepend_unwinder (gdbarch, &amd64_epilogue_frame_unwind);

  /* Hook the prologue-based frame unwinders.  */
  frame_unwind_append_unwinder (gdbarch, &amd64_sigtramp_frame_unwind);
  frame_unwind_append_unwinder (gdbarch, &amd64_frame_unwind);
  frame_base_set_default (gdbarch, &amd64_frame_base);

  set_gdbarch_get_longjmp_target (gdbarch, amd64_get_longjmp_target);

  set_gdbarch_relocate_instruction (gdbarch, amd64_relocate_instruction);

  set_gdbarch_gen_return_address (gdbarch, amd64_gen_return_address);

  /* SystemTap variables and functions.  */
  set_gdbarch_stap_integer_prefixes (gdbarch, stap_integer_prefixes);
  set_gdbarch_stap_register_prefixes (gdbarch, stap_register_prefixes);
  set_gdbarch_stap_register_indirection_prefixes (gdbarch,
					  stap_register_indirection_prefixes);
  set_gdbarch_stap_register_indirection_suffixes (gdbarch,
					  stap_register_indirection_suffixes);
  set_gdbarch_stap_is_single_operand (gdbarch,
				      i386_stap_is_single_operand);
  set_gdbarch_stap_parse_special_token (gdbarch,
					i386_stap_parse_special_token);
  set_gdbarch_insn_is_call (gdbarch, amd64_insn_is_call);
  set_gdbarch_insn_is_ret (gdbarch, amd64_insn_is_ret);
  set_gdbarch_insn_is_jump (gdbarch, amd64_insn_is_jump);
}


static struct type *
amd64_x32_pseudo_register_type (struct gdbarch *gdbarch, int regnum)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  switch (regnum - tdep->eax_regnum)
    {
    case AMD64_RBP_REGNUM:	/* %ebp */
    case AMD64_RSP_REGNUM:	/* %esp */
      return builtin_type (gdbarch)->builtin_data_ptr;
    case AMD64_RIP_REGNUM:	/* %eip */
      return builtin_type (gdbarch)->builtin_func_ptr;
    }

  return i386_pseudo_register_type (gdbarch, regnum);
}

void
amd64_x32_init_abi (struct gdbarch_info info, struct gdbarch *gdbarch,
		    const target_desc *default_tdesc)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  amd64_init_abi (info, gdbarch, default_tdesc);

  tdep->num_dword_regs = 17;
  set_tdesc_pseudo_register_type (gdbarch, amd64_x32_pseudo_register_type);

  set_gdbarch_long_bit (gdbarch, 32);
  set_gdbarch_ptr_bit (gdbarch, 32);
}

/* Return the target description for a specified XSAVE feature mask.  */

const struct target_desc *
amd64_target_description (uint64_t xcr0)
{
  static target_desc *amd64_tdescs \
    [2/*AVX*/][2/*MPX*/][2/*AVX512*/][2/*PKRU*/] = {};
  target_desc **tdesc;

  tdesc = &amd64_tdescs[(xcr0 & X86_XSTATE_AVX) ? 1 : 0]
    [(xcr0 & X86_XSTATE_MPX) ? 1 : 0]
    [(xcr0 & X86_XSTATE_AVX512) ? 1 : 0]
    [(xcr0 & X86_XSTATE_PKRU) ? 1 : 0];

  if (*tdesc == NULL)
    *tdesc = amd64_create_target_description (xcr0, false, false);

  return *tdesc;
}

void
_initialize_amd64_tdep (void)
{
#if GDB_SELF_TEST
  struct
  {
    const char *xml;
    uint64_t mask;
  } xml_masks[] = {
    { "i386/amd64.xml", X86_XSTATE_SSE_MASK },
    { "i386/amd64-avx.xml", X86_XSTATE_AVX_MASK },
    { "i386/amd64-mpx.xml", X86_XSTATE_MPX_MASK },
    { "i386/amd64-avx-mpx.xml", X86_XSTATE_AVX_MPX_MASK },
    { "i386/amd64-avx-avx512.xml", X86_XSTATE_AVX_AVX512_MASK },
    { "i386/amd64-avx-mpx-avx512-pku.xml",
      X86_XSTATE_AVX_MPX_AVX512_PKU_MASK },
  };

  for (auto &a : xml_masks)
    {
      auto tdesc = amd64_target_description (a.mask);

      selftests::record_xml_tdesc (a.xml, tdesc);
    }
#endif /* GDB_SELF_TEST */
}


/* The 64-bit FXSAVE format differs from the 32-bit format in the
   sense that the instruction pointer and data pointer are simply
   64-bit offsets into the code segment and the data segment instead
   of a selector offset pair.  The functions below store the upper 32
   bits of these pointers (instead of just the 16-bits of the segment
   selector).  */

/* Fill register REGNUM in REGCACHE with the appropriate
   floating-point or SSE register value from *FXSAVE.  If REGNUM is
   -1, do this for all registers.  This function masks off any of the
   reserved bits in *FXSAVE.  */

void
amd64_supply_fxsave (struct regcache *regcache, int regnum,
		     const void *fxsave)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  i387_supply_fxsave (regcache, regnum, fxsave);

  if (fxsave
      && gdbarch_bfd_arch_info (gdbarch)->bits_per_word == 64)
    {
      const gdb_byte *regs = (const gdb_byte *) fxsave;

      if (regnum == -1 || regnum == I387_FISEG_REGNUM (tdep))
	regcache_raw_supply (regcache, I387_FISEG_REGNUM (tdep), regs + 12);
      if (regnum == -1 || regnum == I387_FOSEG_REGNUM (tdep))
	regcache_raw_supply (regcache, I387_FOSEG_REGNUM (tdep), regs + 20);
    }
}

/* Similar to amd64_supply_fxsave, but use XSAVE extended state.  */

void
amd64_supply_xsave (struct regcache *regcache, int regnum,
		    const void *xsave)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  i387_supply_xsave (regcache, regnum, xsave);

  if (xsave
      && gdbarch_bfd_arch_info (gdbarch)->bits_per_word == 64)
    {
      const gdb_byte *regs = (const gdb_byte *) xsave;

      if (regnum == -1 || regnum == I387_FISEG_REGNUM (tdep))
	regcache_raw_supply (regcache, I387_FISEG_REGNUM (tdep),
			     regs + 12);
      if (regnum == -1 || regnum == I387_FOSEG_REGNUM (tdep))
	regcache_raw_supply (regcache, I387_FOSEG_REGNUM (tdep),
			     regs + 20);
    }
}

/* Fill register REGNUM (if it is a floating-point or SSE register) in
   *FXSAVE with the value from REGCACHE.  If REGNUM is -1, do this for
   all registers.  This function doesn't touch any of the reserved
   bits in *FXSAVE.  */

void
amd64_collect_fxsave (const struct regcache *regcache, int regnum,
		      void *fxsave)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  gdb_byte *regs = (gdb_byte *) fxsave;

  i387_collect_fxsave (regcache, regnum, fxsave);

  if (gdbarch_bfd_arch_info (gdbarch)->bits_per_word == 64)
    {
      if (regnum == -1 || regnum == I387_FISEG_REGNUM (tdep))
	regcache_raw_collect (regcache, I387_FISEG_REGNUM (tdep), regs + 12);
      if (regnum == -1 || regnum == I387_FOSEG_REGNUM (tdep))
	regcache_raw_collect (regcache, I387_FOSEG_REGNUM (tdep), regs + 20);
    }
}

/* Similar to amd64_collect_fxsave, but use XSAVE extended state.  */

void
amd64_collect_xsave (const struct regcache *regcache, int regnum,
		     void *xsave, int gcore)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  gdb_byte *regs = (gdb_byte *) xsave;

  i387_collect_xsave (regcache, regnum, xsave, gcore);

  if (gdbarch_bfd_arch_info (gdbarch)->bits_per_word == 64)
    {
      if (regnum == -1 || regnum == I387_FISEG_REGNUM (tdep))
	regcache_raw_collect (regcache, I387_FISEG_REGNUM (tdep),
			      regs + 12);
      if (regnum == -1 || regnum == I387_FOSEG_REGNUM (tdep))
	regcache_raw_collect (regcache, I387_FOSEG_REGNUM (tdep),
			      regs + 20);
    }
}