summaryrefslogtreecommitdiff
path: root/gcc/tree-vect-slp.c
blob: 2b075dfea96f773c966709775316da8e054abb46 (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
/* SLP - Basic Block Vectorization
   Copyright (C) 2007-2014 Free Software Foundation, Inc.
   Contributed by Dorit Naishlos <dorit@il.ibm.com>
   and Ira Rosen <irar@il.ibm.com>

This file is part of GCC.

GCC 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, or (at your option) any later
version.

GCC 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 GCC; see the file COPYING3.  If not see
<http://www.gnu.org/licenses/>.  */

#include "config.h"
#include "system.h"
#include "coretypes.h"
#include "dumpfile.h"
#include "tm.h"
#include "tree.h"
#include "stor-layout.h"
#include "target.h"
#include "basic-block.h"
#include "gimple-pretty-print.h"
#include "tree-ssa-alias.h"
#include "internal-fn.h"
#include "gimple-expr.h"
#include "is-a.h"
#include "gimple.h"
#include "gimple-iterator.h"
#include "gimple-ssa.h"
#include "tree-phinodes.h"
#include "ssa-iterators.h"
#include "stringpool.h"
#include "tree-ssanames.h"
#include "tree-pass.h"
#include "cfgloop.h"
#include "expr.h"
#include "recog.h"		/* FIXME: for insn_data */
#include "optabs.h"
#include "tree-vectorizer.h"
#include "langhooks.h"

/* Extract the location of the basic block in the source code.
   Return the basic block location if succeed and NULL if not.  */

source_location
find_bb_location (basic_block bb)
{
  gimple stmt = NULL;
  gimple_stmt_iterator si;

  if (!bb)
    return UNKNOWN_LOCATION;

  for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
    {
      stmt = gsi_stmt (si);
      if (gimple_location (stmt) != UNKNOWN_LOCATION)
        return gimple_location (stmt);
    }

  return UNKNOWN_LOCATION;
}


/* Recursively free the memory allocated for the SLP tree rooted at NODE.  */

static void
vect_free_slp_tree (slp_tree node)
{
  int i;
  slp_tree child;

  if (!node)
    return;

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_free_slp_tree (child);

  SLP_TREE_CHILDREN (node).release ();
  SLP_TREE_SCALAR_STMTS (node).release ();
  SLP_TREE_VEC_STMTS (node).release ();
  SLP_TREE_LOAD_PERMUTATION (node).release ();

  free (node);
}


/* Free the memory allocated for the SLP instance.  */

void
vect_free_slp_instance (slp_instance instance)
{
  vect_free_slp_tree (SLP_INSTANCE_TREE (instance));
  SLP_INSTANCE_LOADS (instance).release ();
  SLP_INSTANCE_BODY_COST_VEC (instance).release ();
  free (instance);
}


/* Create an SLP node for SCALAR_STMTS.  */

static slp_tree
vect_create_new_slp_node (vec<gimple> scalar_stmts)
{
  slp_tree node;
  gimple stmt = scalar_stmts[0];
  unsigned int nops;

  if (is_gimple_call (stmt))
    nops = gimple_call_num_args (stmt);
  else if (is_gimple_assign (stmt))
    {
      nops = gimple_num_ops (stmt) - 1;
      if (gimple_assign_rhs_code (stmt) == COND_EXPR)
	nops++;
    }
  else
    return NULL;

  node = XNEW (struct _slp_tree);
  SLP_TREE_SCALAR_STMTS (node) = scalar_stmts;
  SLP_TREE_VEC_STMTS (node).create (0);
  SLP_TREE_CHILDREN (node).create (nops);
  SLP_TREE_LOAD_PERMUTATION (node) = vNULL;

  return node;
}


/* Allocate operands info for NOPS operands, and GROUP_SIZE def-stmts for each
   operand.  */
static vec<slp_oprnd_info> 
vect_create_oprnd_info (int nops, int group_size)
{
  int i;
  slp_oprnd_info oprnd_info;
  vec<slp_oprnd_info> oprnds_info;

  oprnds_info.create (nops);
  for (i = 0; i < nops; i++)
    {
      oprnd_info = XNEW (struct _slp_oprnd_info);
      oprnd_info->def_stmts.create (group_size);
      oprnd_info->first_dt = vect_uninitialized_def;
      oprnd_info->first_op_type = NULL_TREE;
      oprnd_info->first_pattern = false;
      oprnds_info.quick_push (oprnd_info);
    }

  return oprnds_info;
}


/* Free operands info.  */

static void
vect_free_oprnd_info (vec<slp_oprnd_info> &oprnds_info)
{
  int i;
  slp_oprnd_info oprnd_info;

  FOR_EACH_VEC_ELT (oprnds_info, i, oprnd_info)
    {
      oprnd_info->def_stmts.release ();
      XDELETE (oprnd_info);
    }

  oprnds_info.release ();
}


/* Find the place of the data-ref in STMT in the interleaving chain that starts
   from FIRST_STMT.  Return -1 if the data-ref is not a part of the chain.  */

static int
vect_get_place_in_interleaving_chain (gimple stmt, gimple first_stmt)
{
  gimple next_stmt = first_stmt;
  int result = 0;

  if (first_stmt != GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)))
    return -1;

  do
    {
      if (next_stmt == stmt)
	return result;
      result++;
      next_stmt = GROUP_NEXT_ELEMENT (vinfo_for_stmt (next_stmt));
    }
  while (next_stmt);

  return -1;
}


/* Get the defs for the rhs of STMT (collect them in OPRNDS_INFO), check that
   they are of a valid type and that they match the defs of the first stmt of
   the SLP group (stored in OPRNDS_INFO).  */

static bool
vect_get_and_check_slp_defs (loop_vec_info loop_vinfo, bb_vec_info bb_vinfo,
                             gimple stmt, bool first,
                             vec<slp_oprnd_info> *oprnds_info)
{
  tree oprnd;
  unsigned int i, number_of_oprnds;
  tree def;
  gimple def_stmt;
  enum vect_def_type dt = vect_uninitialized_def;
  struct loop *loop = NULL;
  bool pattern = false;
  slp_oprnd_info oprnd_info;
  int op_idx = 1;
  tree compare_rhs = NULL_TREE;

  if (loop_vinfo)
    loop = LOOP_VINFO_LOOP (loop_vinfo);

  if (is_gimple_call (stmt))
    {
      number_of_oprnds = gimple_call_num_args (stmt);
      op_idx = 3;
    }
  else if (is_gimple_assign (stmt))
    {
      number_of_oprnds = gimple_num_ops (stmt) - 1;
      if (gimple_assign_rhs_code (stmt) == COND_EXPR)
        number_of_oprnds++;
    }
  else
    return false;

  for (i = 0; i < number_of_oprnds; i++)
    {
      if (compare_rhs)
	{
	  oprnd = compare_rhs;
	  compare_rhs = NULL_TREE;
	}
      else
        oprnd = gimple_op (stmt, op_idx++);

      oprnd_info = (*oprnds_info)[i];

      if (COMPARISON_CLASS_P (oprnd))
        {
          compare_rhs = TREE_OPERAND (oprnd, 1);
          oprnd = TREE_OPERAND (oprnd, 0);
	}

      if (!vect_is_simple_use (oprnd, NULL, loop_vinfo, bb_vinfo, &def_stmt,
			       &def, &dt)
	  || (!def_stmt && dt != vect_constant_def))
	{
	  if (dump_enabled_p ())
	    {
	      dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			       "Build SLP failed: can't find def for ");
	      dump_generic_expr (MSG_MISSED_OPTIMIZATION, TDF_SLIM, oprnd);
              dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
	    }

	  return false;
	}

      /* Check if DEF_STMT is a part of a pattern in LOOP and get the def stmt
         from the pattern.  Check that all the stmts of the node are in the
         pattern.  */
      if (def_stmt && gimple_bb (def_stmt)
          && ((loop && flow_bb_inside_loop_p (loop, gimple_bb (def_stmt)))
	      || (!loop && gimple_bb (def_stmt) == BB_VINFO_BB (bb_vinfo)
		  && gimple_code (def_stmt) != GIMPLE_PHI))
          && vinfo_for_stmt (def_stmt)
          && STMT_VINFO_IN_PATTERN_P (vinfo_for_stmt (def_stmt))
	  && !STMT_VINFO_RELEVANT (vinfo_for_stmt (def_stmt))
	  && !STMT_VINFO_LIVE_P (vinfo_for_stmt (def_stmt)))
        {
          pattern = true;
          if (!first && !oprnd_info->first_pattern)
	    {
	      if (dump_enabled_p ())
		{
		  dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				   "Build SLP failed: some of the stmts"
				   " are in a pattern, and others are not ");
		  dump_generic_expr (MSG_MISSED_OPTIMIZATION, TDF_SLIM, oprnd);
                  dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
		}

	      return false;
            }

          def_stmt = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (def_stmt));
          dt = STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def_stmt));

          if (dt == vect_unknown_def_type)
            {
              if (dump_enabled_p ())
                dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				 "Unsupported pattern.\n");
              return false;
            }

          switch (gimple_code (def_stmt))
            {
              case GIMPLE_PHI:
                def = gimple_phi_result (def_stmt);
                break;

              case GIMPLE_ASSIGN:
                def = gimple_assign_lhs (def_stmt);
                break;

              default:
                if (dump_enabled_p ())
                  dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				   "unsupported defining stmt:\n");
                return false;
            }
        }

      if (first)
	{
	  oprnd_info->first_dt = dt;
	  oprnd_info->first_pattern = pattern;
	  oprnd_info->first_op_type = TREE_TYPE (oprnd);
	}
      else
	{
	  /* Not first stmt of the group, check that the def-stmt/s match
	     the def-stmt/s of the first stmt.  Allow different definition
	     types for reduction chains: the first stmt must be a
	     vect_reduction_def (a phi node), and the rest
	     vect_internal_def.  */
	  if (((oprnd_info->first_dt != dt
                && !(oprnd_info->first_dt == vect_reduction_def
                     && dt == vect_internal_def)
		&& !((oprnd_info->first_dt == vect_external_def
		      || oprnd_info->first_dt == vect_constant_def)
		     && (dt == vect_external_def
			 || dt == vect_constant_def)))
               || !types_compatible_p (oprnd_info->first_op_type,
				       TREE_TYPE (oprnd))))
	    {
	      if (dump_enabled_p ())
		dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				 "Build SLP failed: different types\n");

	      return false;
	    }
	}

      /* Check the types of the definitions.  */
      switch (dt)
	{
	case vect_constant_def:
	case vect_external_def:
        case vect_reduction_def:
	  break;

	case vect_internal_def:
	  oprnd_info->def_stmts.quick_push (def_stmt);
	  break;

	default:
	  /* FORNOW: Not supported.  */
	  if (dump_enabled_p ())
	    {
	      dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			       "Build SLP failed: illegal type of def ");
	      dump_generic_expr (MSG_MISSED_OPTIMIZATION, TDF_SLIM, def);
              dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
	    }

	  return false;
	}
    }

  return true;
}


/* Verify if the scalar stmts STMTS are isomorphic, require data
   permutation or are of unsupported types of operation.  Return
   true if they are, otherwise return false and indicate in *MATCHES
   which stmts are not isomorphic to the first one.  If MATCHES[0]
   is false then this indicates the comparison could not be
   carried out or the stmts will never be vectorized by SLP.  */

static bool
vect_build_slp_tree_1 (loop_vec_info loop_vinfo, bb_vec_info bb_vinfo,
		       vec<gimple> stmts, unsigned int group_size,
		       unsigned nops, unsigned int *max_nunits,
		       unsigned int vectorization_factor, bool *matches)
{
  unsigned int i;
  gimple stmt = stmts[0];
  enum tree_code first_stmt_code = ERROR_MARK, rhs_code = ERROR_MARK;
  enum tree_code first_cond_code = ERROR_MARK;
  tree lhs;
  bool need_same_oprnds = false;
  tree vectype, scalar_type, first_op1 = NULL_TREE;
  optab optab;
  int icode;
  enum machine_mode optab_op2_mode;
  enum machine_mode vec_mode;
  struct data_reference *first_dr;
  HOST_WIDE_INT dummy;
  gimple first_load = NULL, prev_first_load = NULL, old_first_load = NULL;
  tree cond;

  /* For every stmt in NODE find its def stmt/s.  */
  FOR_EACH_VEC_ELT (stmts, i, stmt)
    {
      matches[i] = false;

      if (dump_enabled_p ())
	{
	  dump_printf_loc (MSG_NOTE, vect_location, "Build SLP for ");
	  dump_gimple_stmt (MSG_NOTE, TDF_SLIM, stmt, 0);
          dump_printf (MSG_NOTE, "\n");
	}

      /* Fail to vectorize statements marked as unvectorizable.  */
      if (!STMT_VINFO_VECTORIZABLE (vinfo_for_stmt (stmt)))
        {
          if (dump_enabled_p ())
            {
              dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			       "Build SLP failed: unvectorizable statement ");
              dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
              dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
            }
	  /* Fatal mismatch.  */
	  matches[0] = false;
          return false;
        }

      lhs = gimple_get_lhs (stmt);
      if (lhs == NULL_TREE)
	{
	  if (dump_enabled_p ())
	    {
	      dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			       "Build SLP failed: not GIMPLE_ASSIGN nor "
			       "GIMPLE_CALL ");
	      dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
              dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
	    }
	  /* Fatal mismatch.  */
	  matches[0] = false;
	  return false;
	}

       if (is_gimple_assign (stmt)
	   && gimple_assign_rhs_code (stmt) == COND_EXPR
           && (cond = gimple_assign_rhs1 (stmt))
           && !COMPARISON_CLASS_P (cond))
        {
          if (dump_enabled_p ())
            {
              dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location, 
			       "Build SLP failed: condition is not "
			       "comparison ");
              dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
              dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
            }
	  /* Fatal mismatch.  */
	  matches[0] = false;
          return false;
        }

      scalar_type = vect_get_smallest_scalar_type (stmt, &dummy, &dummy);
      vectype = get_vectype_for_scalar_type (scalar_type);
      if (!vectype)
        {
          if (dump_enabled_p ())
            {
              dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location, 
			       "Build SLP failed: unsupported data-type ");
              dump_generic_expr (MSG_MISSED_OPTIMIZATION, TDF_SLIM,
				 scalar_type);
              dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
            }
	  /* Fatal mismatch.  */
	  matches[0] = false;
          return false;
        }

      /* In case of multiple types we need to detect the smallest type.  */
      if (*max_nunits < TYPE_VECTOR_SUBPARTS (vectype))
        {
          *max_nunits = TYPE_VECTOR_SUBPARTS (vectype);
          if (bb_vinfo)
            vectorization_factor = *max_nunits;
        }

      if (is_gimple_call (stmt))
	{
	  rhs_code = CALL_EXPR;
	  if (gimple_call_internal_p (stmt)
	      || gimple_call_tail_p (stmt)
	      || gimple_call_noreturn_p (stmt)
	      || !gimple_call_nothrow_p (stmt)
	      || gimple_call_chain (stmt))
	    {
	      if (dump_enabled_p ())
		{
		  dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location, 
				   "Build SLP failed: unsupported call type ");
		  dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
                  dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
		}
	      /* Fatal mismatch.  */
	      matches[0] = false;
	      return false;
	    }
	}
      else
	rhs_code = gimple_assign_rhs_code (stmt);

      /* Check the operation.  */
      if (i == 0)
	{
	  first_stmt_code = rhs_code;

	  /* Shift arguments should be equal in all the packed stmts for a
	     vector shift with scalar shift operand.  */
	  if (rhs_code == LSHIFT_EXPR || rhs_code == RSHIFT_EXPR
	      || rhs_code == LROTATE_EXPR
	      || rhs_code == RROTATE_EXPR)
	    {
	      vec_mode = TYPE_MODE (vectype);

	      /* First see if we have a vector/vector shift.  */
	      optab = optab_for_tree_code (rhs_code, vectype,
					   optab_vector);

	      if (!optab
		  || optab_handler (optab, vec_mode) == CODE_FOR_nothing)
		{
		  /* No vector/vector shift, try for a vector/scalar shift.  */
		  optab = optab_for_tree_code (rhs_code, vectype,
					       optab_scalar);

		  if (!optab)
		    {
		      if (dump_enabled_p ())
			dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
					 "Build SLP failed: no optab.\n");
		      /* Fatal mismatch.  */
		      matches[0] = false;
		      return false;
		    }
		  icode = (int) optab_handler (optab, vec_mode);
		  if (icode == CODE_FOR_nothing)
		    {
		      if (dump_enabled_p ())
			dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
					 "Build SLP failed: "
					 "op not supported by target.\n");
		      /* Fatal mismatch.  */
		      matches[0] = false;
		      return false;
		    }
		  optab_op2_mode = insn_data[icode].operand[2].mode;
		  if (!VECTOR_MODE_P (optab_op2_mode))
		    {
		      need_same_oprnds = true;
		      first_op1 = gimple_assign_rhs2 (stmt);
		    }
		}
	    }
	  else if (rhs_code == WIDEN_LSHIFT_EXPR)
            {
              need_same_oprnds = true;
              first_op1 = gimple_assign_rhs2 (stmt);
            }
	}
      else
	{
	  if (first_stmt_code != rhs_code
	      && (first_stmt_code != IMAGPART_EXPR
		  || rhs_code != REALPART_EXPR)
	      && (first_stmt_code != REALPART_EXPR
		  || rhs_code != IMAGPART_EXPR)
              && !(STMT_VINFO_GROUPED_ACCESS (vinfo_for_stmt (stmt))
                   && (first_stmt_code == ARRAY_REF
                       || first_stmt_code == BIT_FIELD_REF
                       || first_stmt_code == INDIRECT_REF
                       || first_stmt_code == COMPONENT_REF
                       || first_stmt_code == MEM_REF)))
	    {
	      if (dump_enabled_p ())
		{
		  dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location, 
				   "Build SLP failed: different operation "
				   "in stmt ");
		  dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
                  dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
		}
	      /* Mismatch.  */
	      continue;
	    }

	  if (need_same_oprnds
	      && !operand_equal_p (first_op1, gimple_assign_rhs2 (stmt), 0))
	    {
	      if (dump_enabled_p ())
		{
		  dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location, 
				   "Build SLP failed: different shift "
				   "arguments in ");
		  dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
                  dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
		}
	      /* Mismatch.  */
	      continue;
	    }

	  if (rhs_code == CALL_EXPR)
	    {
	      gimple first_stmt = stmts[0];
	      if (gimple_call_num_args (stmt) != nops
		  || !operand_equal_p (gimple_call_fn (first_stmt),
				       gimple_call_fn (stmt), 0)
		  || gimple_call_fntype (first_stmt)
		     != gimple_call_fntype (stmt))
		{
		  if (dump_enabled_p ())
		    {
		      dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location, 
				       "Build SLP failed: different calls in ");
		      dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM,
					stmt, 0);
                      dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
		    }
		  /* Mismatch.  */
		  continue;
		}
	    }
	}

      /* Grouped store or load.  */
      if (STMT_VINFO_GROUPED_ACCESS (vinfo_for_stmt (stmt)))
	{
	  if (REFERENCE_CLASS_P (lhs))
	    {
	      /* Store.  */
	      ;
	    }
	  else
	    {
	      /* Load.  */
	      unsigned unrolling_factor
		= least_common_multiple
		    (*max_nunits, group_size) / group_size;
              /* FORNOW: Check that there is no gap between the loads
		 and no gap between the groups when we need to load
		 multiple groups at once.
		 ???  We should enhance this to only disallow gaps
		 inside vectors.  */
              if ((unrolling_factor > 1
		   && GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)) == stmt
		   && GROUP_GAP (vinfo_for_stmt (stmt)) != 0)
		  || (GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)) != stmt
		      && GROUP_GAP (vinfo_for_stmt (stmt)) != 1))
                {
                  if (dump_enabled_p ())
                    {
                      dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				       "Build SLP failed: grouped "
				       "loads have gaps ");
                      dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM,
					stmt, 0);
                      dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
                    }
		  /* Fatal mismatch.  */
		  matches[0] = false;
                  return false;
                }

              /* Check that the size of interleaved loads group is not
                 greater than the SLP group size.  */
	      unsigned ncopies
		= vectorization_factor / TYPE_VECTOR_SUBPARTS (vectype);
              if (loop_vinfo
		  && GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)) == stmt
                  && ((GROUP_SIZE (vinfo_for_stmt (stmt))
		       - GROUP_GAP (vinfo_for_stmt (stmt)))
		      > ncopies * group_size))
                {
                  if (dump_enabled_p ())
                    {
                      dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				       "Build SLP failed: the number "
				       "of interleaved loads is greater than "
				       "the SLP group size ");
                      dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM,
					stmt, 0);
                      dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
                    }
		  /* Fatal mismatch.  */
		  matches[0] = false;
                  return false;
                }

	      old_first_load = first_load;
              first_load = GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt));
              if (prev_first_load)
                {
                  /* Check that there are no loads from different interleaving
                     chains in the same node.  */
                  if (prev_first_load != first_load)
                    {
                      if (dump_enabled_p ())
                        {
                          dump_printf_loc (MSG_MISSED_OPTIMIZATION,
					   vect_location, 
					   "Build SLP failed: different "
					   "interleaving chains in one node ");
                          dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM,
					    stmt, 0);
                          dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
                        }
		      /* Mismatch.  */
		      continue;
                    }
                }
              else
                prev_first_load = first_load;

	      /* In some cases a group of loads is just the same load
		 repeated N times.  Only analyze its cost once.  */
              if (first_load == stmt && old_first_load != first_load)
                {
                  first_dr = STMT_VINFO_DATA_REF (vinfo_for_stmt (stmt));
                  if (vect_supportable_dr_alignment (first_dr, false)
                      == dr_unaligned_unsupported)
                    {
                      if (dump_enabled_p ())
                        {
                          dump_printf_loc (MSG_MISSED_OPTIMIZATION,
					   vect_location, 
					   "Build SLP failed: unsupported "
					   "unaligned load ");
                          dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM,
					    stmt, 0);
                          dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
                        }
		      /* Fatal mismatch.  */
		      matches[0] = false;
                      return false;
                    }
                }
           }
        } /* Grouped access.  */
      else
	{
	  if (TREE_CODE_CLASS (rhs_code) == tcc_reference)
	    {
	      /* Not grouped load.  */
	      if (dump_enabled_p ())
		{
		  dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location, 
				   "Build SLP failed: not grouped load ");
		  dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
                  dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
		}

	      /* FORNOW: Not grouped loads are not supported.  */
	      /* Fatal mismatch.  */
	      matches[0] = false;
	      return false;
	    }

	  /* Not memory operation.  */
	  if (TREE_CODE_CLASS (rhs_code) != tcc_binary
	      && TREE_CODE_CLASS (rhs_code) != tcc_unary
	      && rhs_code != COND_EXPR
	      && rhs_code != CALL_EXPR)
	    {
	      if (dump_enabled_p ())
		{
		  dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				   "Build SLP failed: operation");
		  dump_printf (MSG_MISSED_OPTIMIZATION, " unsupported ");
		  dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
                  dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
		}
	      /* Fatal mismatch.  */
	      matches[0] = false;
	      return false;
	    }

          if (rhs_code == COND_EXPR)
            {
              tree cond_expr = gimple_assign_rhs1 (stmt);

	      if (i == 0)
		first_cond_code = TREE_CODE (cond_expr);
              else if (first_cond_code != TREE_CODE (cond_expr))
                {
                  if (dump_enabled_p ())
                    {
                      dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				       "Build SLP failed: different"
				       " operation");
                      dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM,
					stmt, 0);
                      dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
                    }
		  /* Mismatch.  */
		  continue;
		}
            }
	}

      matches[i] = true;
    }

  for (i = 0; i < group_size; ++i)
    if (!matches[i])
      return false;

  return true;
}

/* Recursively build an SLP tree starting from NODE.
   Fail (and return a value not equal to zero) if def-stmts are not
   isomorphic, require data permutation or are of unsupported types of
   operation.  Otherwise, return 0.
   The value returned is the depth in the SLP tree where a mismatch
   was found.  */

static bool
vect_build_slp_tree (loop_vec_info loop_vinfo, bb_vec_info bb_vinfo,
                     slp_tree *node, unsigned int group_size,
                     unsigned int *max_nunits,
                     vec<slp_tree> *loads,
                     unsigned int vectorization_factor,
		     bool *matches, unsigned *npermutes)
{
  unsigned nops, i, this_npermutes = 0;
  gimple stmt;

  if (!matches)
    matches = XALLOCAVEC (bool, group_size);
  if (!npermutes)
    npermutes = &this_npermutes;

  matches[0] = false;

  stmt = SLP_TREE_SCALAR_STMTS (*node)[0];
  if (is_gimple_call (stmt))
    nops = gimple_call_num_args (stmt);
  else if (is_gimple_assign (stmt))
    {
      nops = gimple_num_ops (stmt) - 1;
      if (gimple_assign_rhs_code (stmt) == COND_EXPR)
	nops++;
    }
  else
    return false;

  if (!vect_build_slp_tree_1 (loop_vinfo, bb_vinfo,
			      SLP_TREE_SCALAR_STMTS (*node), group_size, nops,
			      max_nunits, vectorization_factor, matches))
    return false;

  /* If the SLP node is a load, terminate the recursion.  */
  if (STMT_VINFO_GROUPED_ACCESS (vinfo_for_stmt (stmt))
      && DR_IS_READ (STMT_VINFO_DATA_REF (vinfo_for_stmt (stmt))))
    {
      loads->safe_push (*node);
      return true;
    }

  /* Get at the operands, verifying they are compatible.  */
  vec<slp_oprnd_info> oprnds_info = vect_create_oprnd_info (nops, group_size);
  slp_oprnd_info oprnd_info;
  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (*node), i, stmt)
    {
      if (!vect_get_and_check_slp_defs (loop_vinfo, bb_vinfo,
					stmt, (i == 0), &oprnds_info))
	{
	  vect_free_oprnd_info (oprnds_info);
	  return false;
	}
    }

  stmt = SLP_TREE_SCALAR_STMTS (*node)[0];

  /* Create SLP_TREE nodes for the definition node/s.  */
  FOR_EACH_VEC_ELT (oprnds_info, i, oprnd_info)
    {
      slp_tree child;
      unsigned old_nloads = loads->length ();
      unsigned old_max_nunits = *max_nunits;

      if (oprnd_info->first_dt != vect_internal_def)
        continue;

      child = vect_create_new_slp_node (oprnd_info->def_stmts);
      if (!child)
	{
	  vect_free_oprnd_info (oprnds_info);
	  return false;
	}

      bool *matches = XALLOCAVEC (bool, group_size);
      if (vect_build_slp_tree (loop_vinfo, bb_vinfo, &child,
			       group_size, max_nunits, loads,
			       vectorization_factor, matches, npermutes))
	{
	  oprnd_info->def_stmts = vNULL;
	  SLP_TREE_CHILDREN (*node).quick_push (child);
	  continue;
	}

      /* If the SLP build for operand zero failed and operand zero
	 and one can be commutated try that for the scalar stmts
	 that failed the match.  */
      if (i == 0
	  /* A first scalar stmt mismatch signals a fatal mismatch.  */
	  && matches[0]
	  /* ???  For COND_EXPRs we can swap the comparison operands
	     as well as the arms under some constraints.  */
	  && nops == 2
	  && oprnds_info[1]->first_dt == vect_internal_def
	  && is_gimple_assign (stmt)
	  && commutative_tree_code (gimple_assign_rhs_code (stmt))
	  /* Do so only if the number of not successful permutes was nor more
	     than a cut-ff as re-trying the recursive match on
	     possibly each level of the tree would expose exponential
	     behavior.  */
	  && *npermutes < 4)
	{
	  /* Roll back.  */
	  *max_nunits = old_max_nunits;
	  loads->truncate (old_nloads);
	  /* Swap mismatched definition stmts.  */
	  for (unsigned j = 0; j < group_size; ++j)
	    if (!matches[j])
	      {
		gimple tem = oprnds_info[0]->def_stmts[j];
		oprnds_info[0]->def_stmts[j] = oprnds_info[1]->def_stmts[j];
		oprnds_info[1]->def_stmts[j] = tem;
	      }
	  /* And try again ... */
	  if (vect_build_slp_tree (loop_vinfo, bb_vinfo, &child,
				   group_size, max_nunits, loads,
				   vectorization_factor,
				   matches, npermutes))
	    {
	      oprnd_info->def_stmts = vNULL;
	      SLP_TREE_CHILDREN (*node).quick_push (child);
	      continue;
	    }

	  ++*npermutes;
	}

      oprnd_info->def_stmts = vNULL;
      vect_free_slp_tree (child);
      vect_free_oprnd_info (oprnds_info);
      return false;
    }

  vect_free_oprnd_info (oprnds_info);
  return true;
}

/* Dump a slp tree NODE using flags specified in DUMP_KIND.  */

static void
vect_print_slp_tree (int dump_kind, slp_tree node)
{
  int i;
  gimple stmt;
  slp_tree child;

  if (!node)
    return;

  dump_printf (dump_kind, "node ");
  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, stmt)
    {
      dump_printf (dump_kind, "\n\tstmt %d ", i);
      dump_gimple_stmt (dump_kind, TDF_SLIM, stmt, 0);
    }
  dump_printf (dump_kind, "\n");

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_print_slp_tree (dump_kind, child);
}


/* Mark the tree rooted at NODE with MARK (PURE_SLP or HYBRID).
   If MARK is HYBRID, it refers to a specific stmt in NODE (the stmt at index
   J).  Otherwise, MARK is PURE_SLP and J is -1, which indicates that all the
   stmts in NODE are to be marked.  */

static void
vect_mark_slp_stmts (slp_tree node, enum slp_vect_type mark, int j)
{
  int i;
  gimple stmt;
  slp_tree child;

  if (!node)
    return;

  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, stmt)
    if (j < 0 || i == j)
      STMT_SLP_TYPE (vinfo_for_stmt (stmt)) = mark;

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_mark_slp_stmts (child, mark, j);
}


/* Mark the statements of the tree rooted at NODE as relevant (vect_used).  */

static void
vect_mark_slp_stmts_relevant (slp_tree node)
{
  int i;
  gimple stmt;
  stmt_vec_info stmt_info;
  slp_tree child;

  if (!node)
    return;

  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, stmt)
    {
      stmt_info = vinfo_for_stmt (stmt);
      gcc_assert (!STMT_VINFO_RELEVANT (stmt_info)
                  || STMT_VINFO_RELEVANT (stmt_info) == vect_used_in_scope);
      STMT_VINFO_RELEVANT (stmt_info) = vect_used_in_scope;
    }

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_mark_slp_stmts_relevant (child);
}


/* Rearrange the statements of NODE according to PERMUTATION.  */

static void
vect_slp_rearrange_stmts (slp_tree node, unsigned int group_size,
                          vec<unsigned> permutation)
{
  gimple stmt;
  vec<gimple> tmp_stmts;
  unsigned int i;
  slp_tree child;

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_slp_rearrange_stmts (child, group_size, permutation);

  gcc_assert (group_size == SLP_TREE_SCALAR_STMTS (node).length ());
  tmp_stmts.create (group_size);
  tmp_stmts.quick_grow_cleared (group_size);

  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, stmt)
    tmp_stmts[permutation[i]] = stmt;

  SLP_TREE_SCALAR_STMTS (node).release ();
  SLP_TREE_SCALAR_STMTS (node) = tmp_stmts;
}


/* Check if the required load permutations in the SLP instance
   SLP_INSTN are supported.  */

static bool
vect_supported_load_permutation_p (slp_instance slp_instn)
{
  unsigned int group_size = SLP_INSTANCE_GROUP_SIZE (slp_instn);
  unsigned int i, j, k, next;
  sbitmap load_index;
  slp_tree node;
  gimple stmt, load, next_load, first_load;
  struct data_reference *dr;

  if (dump_enabled_p ())
    {
      dump_printf_loc (MSG_NOTE, vect_location, "Load permutation ");
      FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
	if (node->load_permutation.exists ())
	  FOR_EACH_VEC_ELT (node->load_permutation, j, next)
	    dump_printf (MSG_NOTE, "%d ", next);
	else
	  for (i = 0; i < group_size; ++i)
	    dump_printf (MSG_NOTE, "%d ", i);
      dump_printf (MSG_NOTE, "\n");
    }

  /* In case of reduction every load permutation is allowed, since the order
     of the reduction statements is not important (as opposed to the case of
     grouped stores).  The only condition we need to check is that all the
     load nodes are of the same size and have the same permutation (and then
     rearrange all the nodes of the SLP instance according to this 
     permutation).  */

  /* Check that all the load nodes are of the same size.  */
  /* ???  Can't we assert this? */
  FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
    if (SLP_TREE_SCALAR_STMTS (node).length () != (unsigned) group_size)
      return false;

  node = SLP_INSTANCE_TREE (slp_instn);
  stmt = SLP_TREE_SCALAR_STMTS (node)[0];

  /* Reduction (there are no data-refs in the root).
     In reduction chain the order of the loads is important.  */
  if (!STMT_VINFO_DATA_REF (vinfo_for_stmt (stmt))
      && !GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)))
    {
      slp_tree load;
      unsigned int lidx;

      /* Compare all the permutation sequences to the first one.  We know
         that at least one load is permuted.  */
      node = SLP_INSTANCE_LOADS (slp_instn)[0];
      if (!node->load_permutation.exists ())
	return false;
      for (i = 1; SLP_INSTANCE_LOADS (slp_instn).iterate (i, &load); ++i)
	{
	  if (!load->load_permutation.exists ())
	    return false;
	  FOR_EACH_VEC_ELT (load->load_permutation, j, lidx)
	    if (lidx != node->load_permutation[j])
	      return false;
	}

      /* Check that the loads in the first sequence are different and there
	 are no gaps between them.  */
      load_index = sbitmap_alloc (group_size);
      bitmap_clear (load_index);
      FOR_EACH_VEC_ELT (node->load_permutation, i, lidx)
	{
	  if (bitmap_bit_p (load_index, lidx))
	    {
	      sbitmap_free (load_index);
	      return false;
	    }
	  bitmap_set_bit (load_index, lidx);
	}
      for (i = 0; i < group_size; i++)
	if (!bitmap_bit_p (load_index, i))
	  {
	    sbitmap_free (load_index);
	    return false;
	  }
      sbitmap_free (load_index);

      /* This permutation is valid for reduction.  Since the order of the
	 statements in the nodes is not important unless they are memory
	 accesses, we can rearrange the statements in all the nodes
	 according to the order of the loads.  */
      vect_slp_rearrange_stmts (SLP_INSTANCE_TREE (slp_instn), group_size,
				node->load_permutation);

      /* We are done, no actual permutations need to be generated.  */
      FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
	SLP_TREE_LOAD_PERMUTATION (node).release ();
      return true;
    }

  /* In basic block vectorization we allow any subchain of an interleaving
     chain.
     FORNOW: not supported in loop SLP because of realignment compications.  */
  if (STMT_VINFO_BB_VINFO (vinfo_for_stmt (stmt)))
    {
      /* Check that for every node in the instance the loads
	 form a subchain.  */
      FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
        {
          next_load = NULL;
          FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), j, load)
            {
              if (j != 0 && next_load != load)
		return false;
              next_load = GROUP_NEXT_ELEMENT (vinfo_for_stmt (load));
            }
        }

      /* Check that the alignment of the first load in every subchain, i.e.,
         the first statement in every load node, is supported.
	 ???  This belongs in alignment checking.  */
      FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
	{
	  first_load = SLP_TREE_SCALAR_STMTS (node)[0];
	  if (first_load != GROUP_FIRST_ELEMENT (vinfo_for_stmt (first_load)))
	    {
	      dr = STMT_VINFO_DATA_REF (vinfo_for_stmt (first_load));
	      if (vect_supportable_dr_alignment (dr, false)
		  == dr_unaligned_unsupported)
		{
		  if (dump_enabled_p ())
		    {
		      dump_printf_loc (MSG_MISSED_OPTIMIZATION,
				       vect_location,
				       "unsupported unaligned load ");
		      dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM,
					first_load, 0);
                      dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
		    }
		  return false;
		}
	    }
	}

      /* We are done, no actual permutations need to be generated.  */
      FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
	SLP_TREE_LOAD_PERMUTATION (node).release ();
      return true;
    }

  /* FORNOW: the only supported permutation is 0..01..1.. of length equal to
     GROUP_SIZE and where each sequence of same drs is of GROUP_SIZE length as
     well (unless it's reduction).  */
  if (SLP_INSTANCE_LOADS (slp_instn).length () != group_size)
    return false;
  FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
    if (!node->load_permutation.exists ())
      return false;

  load_index = sbitmap_alloc (group_size);
  bitmap_clear (load_index);
  FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
    {
      unsigned int lidx = node->load_permutation[0];
      if (bitmap_bit_p (load_index, lidx))
	{
	  sbitmap_free (load_index);
	  return false;
	}
      bitmap_set_bit (load_index, lidx);
      FOR_EACH_VEC_ELT (node->load_permutation, j, k)
	if (k != lidx)
	  {
	    sbitmap_free (load_index);
	    return false;
	  }
    }
  for (i = 0; i < group_size; i++)
    if (!bitmap_bit_p (load_index, i))
      {
	sbitmap_free (load_index);
	return false;
      }
  sbitmap_free (load_index);

  FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (slp_instn), i, node)
    if (node->load_permutation.exists ()
	&& !vect_transform_slp_perm_load
	      (node, vNULL, NULL,
	       SLP_INSTANCE_UNROLLING_FACTOR (slp_instn), slp_instn, true))
      return false;
  return true;
}


/* Find the first load in the loop that belongs to INSTANCE.
   When loads are in several SLP nodes, there can be a case in which the first
   load does not appear in the first SLP node to be transformed, causing
   incorrect order of statements.  Since we generate all the loads together,
   they must be inserted before the first load of the SLP instance and not
   before the first load of the first node of the instance.  */

static gimple
vect_find_first_load_in_slp_instance (slp_instance instance)
{
  int i, j;
  slp_tree load_node;
  gimple first_load = NULL, load;

  FOR_EACH_VEC_ELT (SLP_INSTANCE_LOADS (instance), i, load_node)
    FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (load_node), j, load)
      first_load = get_earlier_stmt (load, first_load);

  return first_load;
}


/* Find the last store in SLP INSTANCE.  */

static gimple
vect_find_last_store_in_slp_instance (slp_instance instance)
{
  int i;
  slp_tree node;
  gimple last_store = NULL, store;

  node = SLP_INSTANCE_TREE (instance);
  for (i = 0; SLP_TREE_SCALAR_STMTS (node).iterate (i, &store); i++)
    last_store = get_later_stmt (store, last_store);

  return last_store;
}

/* Compute the cost for the SLP node NODE in the SLP instance INSTANCE.  */

static void
vect_analyze_slp_cost_1 (loop_vec_info loop_vinfo, bb_vec_info bb_vinfo,
			 slp_instance instance, slp_tree node,
			 stmt_vector_for_cost *prologue_cost_vec,
			 unsigned ncopies_for_cost)
{
  stmt_vector_for_cost *body_cost_vec = &SLP_INSTANCE_BODY_COST_VEC (instance);

  unsigned i;
  slp_tree child;
  gimple stmt, s;
  stmt_vec_info stmt_info;
  tree lhs;
  unsigned group_size = SLP_INSTANCE_GROUP_SIZE (instance);

  /* Recurse down the SLP tree.  */
  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_analyze_slp_cost_1 (loop_vinfo, bb_vinfo,
			     instance, child, prologue_cost_vec,
			     ncopies_for_cost);

  /* Look at the first scalar stmt to determine the cost.  */
  stmt = SLP_TREE_SCALAR_STMTS (node)[0];
  stmt_info = vinfo_for_stmt (stmt);
  if (STMT_VINFO_GROUPED_ACCESS (stmt_info))
    {
      if (DR_IS_WRITE (STMT_VINFO_DATA_REF (stmt_info)))
	vect_model_store_cost (stmt_info, ncopies_for_cost, false,
			       vect_uninitialized_def,
			       node, prologue_cost_vec, body_cost_vec);
      else
	{
	  int i;
	  gcc_checking_assert (DR_IS_READ (STMT_VINFO_DATA_REF (stmt_info)));
	  vect_model_load_cost (stmt_info, ncopies_for_cost, false,
				node, prologue_cost_vec, body_cost_vec);
	  /* If the load is permuted record the cost for the permutation.
	     ???  Loads from multiple chains are let through here only
	     for a single special case involving complex numbers where
	     in the end no permutation is necessary.  */
	  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, s)
	    if ((STMT_VINFO_GROUP_FIRST_ELEMENT (vinfo_for_stmt (s))
		 == STMT_VINFO_GROUP_FIRST_ELEMENT (stmt_info))
		&& vect_get_place_in_interleaving_chain
		     (s, STMT_VINFO_GROUP_FIRST_ELEMENT (stmt_info)) != i)
	      {
		record_stmt_cost (body_cost_vec, group_size, vec_perm,
				  stmt_info, 0, vect_body);
		break;
	      }
	}
    }
  else
    record_stmt_cost (body_cost_vec, ncopies_for_cost, vector_stmt,
		      stmt_info, 0, vect_body);

  /* Scan operands and account for prologue cost of constants/externals.
     ???  This over-estimates cost for multiple uses and should be
     re-engineered.  */
  lhs = gimple_get_lhs (stmt);
  for (i = 0; i < gimple_num_ops (stmt); ++i)
    {
      tree def, op = gimple_op (stmt, i);
      gimple def_stmt;
      enum vect_def_type dt;
      if (!op || op == lhs)
	continue;
      if (vect_is_simple_use (op, NULL, loop_vinfo, bb_vinfo,
			      &def_stmt, &def, &dt)
	  && (dt == vect_constant_def || dt == vect_external_def))
	record_stmt_cost (prologue_cost_vec, 1, vector_stmt,
			  stmt_info, 0, vect_prologue);
    }
}

/* Compute the cost for the SLP instance INSTANCE.  */

static void
vect_analyze_slp_cost (loop_vec_info loop_vinfo, bb_vec_info bb_vinfo,
		       slp_instance instance, unsigned nunits)
{
  stmt_vector_for_cost body_cost_vec, prologue_cost_vec;
  unsigned ncopies_for_cost;
  stmt_info_for_cost *si;
  unsigned i;

  /* Calculate the number of vector stmts to create based on the unrolling
     factor (number of vectors is 1 if NUNITS >= GROUP_SIZE, and is
     GROUP_SIZE / NUNITS otherwise.  */
  unsigned group_size = SLP_INSTANCE_GROUP_SIZE (instance);
  ncopies_for_cost = least_common_multiple (nunits, group_size) / nunits;

  prologue_cost_vec.create (10);
  body_cost_vec.create (10);
  SLP_INSTANCE_BODY_COST_VEC (instance) = body_cost_vec;
  vect_analyze_slp_cost_1 (loop_vinfo, bb_vinfo,
			   instance, SLP_INSTANCE_TREE (instance),
			   &prologue_cost_vec, ncopies_for_cost);

  /* Record the prologue costs, which were delayed until we were
     sure that SLP was successful.  Unlike the body costs, we know
     the final values now regardless of the loop vectorization factor.  */
  void *data = (loop_vinfo ? LOOP_VINFO_TARGET_COST_DATA (loop_vinfo)
		: BB_VINFO_TARGET_COST_DATA (bb_vinfo));
  FOR_EACH_VEC_ELT (prologue_cost_vec, i, si)
    {
      struct _stmt_vec_info *stmt_info
	= si->stmt ? vinfo_for_stmt (si->stmt) : NULL;
      (void) add_stmt_cost (data, si->count, si->kind, stmt_info,
			    si->misalign, vect_prologue);
    }

  prologue_cost_vec.release ();
}

/* Analyze an SLP instance starting from a group of grouped stores.  Call
   vect_build_slp_tree to build a tree of packed stmts if possible.
   Return FALSE if it's impossible to SLP any stmt in the loop.  */

static bool
vect_analyze_slp_instance (loop_vec_info loop_vinfo, bb_vec_info bb_vinfo,
                           gimple stmt)
{
  slp_instance new_instance;
  slp_tree node;
  unsigned int group_size = GROUP_SIZE (vinfo_for_stmt (stmt));
  unsigned int unrolling_factor = 1, nunits;
  tree vectype, scalar_type = NULL_TREE;
  gimple next;
  unsigned int vectorization_factor = 0;
  int i;
  unsigned int max_nunits = 0;
  vec<slp_tree> loads;
  struct data_reference *dr = STMT_VINFO_DATA_REF (vinfo_for_stmt (stmt));
  vec<gimple> scalar_stmts;

  if (GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)))
    {
      if (dr)
        {
          scalar_type = TREE_TYPE (DR_REF (dr));
          vectype = get_vectype_for_scalar_type (scalar_type);
        }
      else
        {
          gcc_assert (loop_vinfo);
          vectype = STMT_VINFO_VECTYPE (vinfo_for_stmt (stmt));
        }

      group_size = GROUP_SIZE (vinfo_for_stmt (stmt));
    }
  else
    {
      gcc_assert (loop_vinfo);
      vectype = STMT_VINFO_VECTYPE (vinfo_for_stmt (stmt));
      group_size = LOOP_VINFO_REDUCTIONS (loop_vinfo).length ();
    }

  if (!vectype)
    {
      if (dump_enabled_p ())
        {
          dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			   "Build SLP failed: unsupported data-type ");
          dump_generic_expr (MSG_MISSED_OPTIMIZATION, TDF_SLIM, scalar_type);
          dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
        }

      return false;
    }

  nunits = TYPE_VECTOR_SUBPARTS (vectype);
  if (loop_vinfo)
    vectorization_factor = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
  else
    vectorization_factor = nunits;

  /* Calculate the unrolling factor.  */
  unrolling_factor = least_common_multiple (nunits, group_size) / group_size;
  if (unrolling_factor != 1 && !loop_vinfo)
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			 "Build SLP failed: unrolling required in basic"
			 " block SLP\n");

      return false;
    }

  /* Create a node (a root of the SLP tree) for the packed grouped stores.  */
  scalar_stmts.create (group_size);
  next = stmt;
  if (GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)))
    {
      /* Collect the stores and store them in SLP_TREE_SCALAR_STMTS.  */
      while (next)
        {
	  if (STMT_VINFO_IN_PATTERN_P (vinfo_for_stmt (next))
	      && STMT_VINFO_RELATED_STMT (vinfo_for_stmt (next)))
	    scalar_stmts.safe_push (
		  STMT_VINFO_RELATED_STMT (vinfo_for_stmt (next)));
	  else
            scalar_stmts.safe_push (next);
          next = GROUP_NEXT_ELEMENT (vinfo_for_stmt (next));
        }
    }
  else
    {
      /* Collect reduction statements.  */
      vec<gimple> reductions = LOOP_VINFO_REDUCTIONS (loop_vinfo);
      for (i = 0; reductions.iterate (i, &next); i++)
	scalar_stmts.safe_push (next);
    }

  node = vect_create_new_slp_node (scalar_stmts);

  loads.create (group_size);

  /* Build the tree for the SLP instance.  */
  if (vect_build_slp_tree (loop_vinfo, bb_vinfo, &node, group_size,
			   &max_nunits, &loads,
			   vectorization_factor, NULL, NULL))
    {
      /* Calculate the unrolling factor based on the smallest type.  */
      if (max_nunits > nunits)
        unrolling_factor = least_common_multiple (max_nunits, group_size)
                           / group_size;

      if (unrolling_factor != 1 && !loop_vinfo)
        {
          if (dump_enabled_p ())
            dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			     "Build SLP failed: unrolling required in basic"
			     " block SLP\n");
	  vect_free_slp_tree (node);
	  loads.release ();
          return false;
        }

      /* Create a new SLP instance.  */
      new_instance = XNEW (struct _slp_instance);
      SLP_INSTANCE_TREE (new_instance) = node;
      SLP_INSTANCE_GROUP_SIZE (new_instance) = group_size;
      SLP_INSTANCE_UNROLLING_FACTOR (new_instance) = unrolling_factor;
      SLP_INSTANCE_BODY_COST_VEC (new_instance) = vNULL;
      SLP_INSTANCE_LOADS (new_instance) = loads;
      SLP_INSTANCE_FIRST_LOAD_STMT (new_instance) = NULL;

      /* Compute the load permutation.  */
      slp_tree load_node;
      bool loads_permuted = false;
      FOR_EACH_VEC_ELT (loads, i, load_node)
	{
	  vec<unsigned> load_permutation;
	  int j;
	  gimple load, first_stmt;
	  bool this_load_permuted = false;
	  load_permutation.create (group_size);
	  first_stmt = GROUP_FIRST_ELEMENT
	      (vinfo_for_stmt (SLP_TREE_SCALAR_STMTS (load_node)[0]));
	  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (load_node), j, load)
	    {
	      int load_place
		= vect_get_place_in_interleaving_chain (load, first_stmt);
	      gcc_assert (load_place != -1);
	      if (load_place != j)
		this_load_permuted = true;
	      load_permutation.safe_push (load_place);
	    }
	  if (!this_load_permuted)
	    {
	      load_permutation.release ();
	      continue;
	    }
	  SLP_TREE_LOAD_PERMUTATION (load_node) = load_permutation;
	  loads_permuted = true;
	}

      if (loads_permuted)
        {
          if (!vect_supported_load_permutation_p (new_instance))
            {
              if (dump_enabled_p ())
                {
                  dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
				   "Build SLP failed: unsupported load "
				   "permutation ");
                  dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
                  dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
                }
              vect_free_slp_instance (new_instance);
              return false;
            }

          SLP_INSTANCE_FIRST_LOAD_STMT (new_instance)
	    = vect_find_first_load_in_slp_instance (new_instance);
        }

      /* Compute the costs of this SLP instance.  */
      vect_analyze_slp_cost (loop_vinfo, bb_vinfo,
			     new_instance, TYPE_VECTOR_SUBPARTS (vectype));

      if (loop_vinfo)
        LOOP_VINFO_SLP_INSTANCES (loop_vinfo).safe_push (new_instance);
      else
        BB_VINFO_SLP_INSTANCES (bb_vinfo).safe_push (new_instance);

      if (dump_enabled_p ())
	vect_print_slp_tree (MSG_NOTE, node);

      return true;
    }

  /* Failed to SLP.  */
  /* Free the allocated memory.  */
  vect_free_slp_tree (node);
  loads.release ();

  return false;
}


/* Check if there are stmts in the loop can be vectorized using SLP.  Build SLP
   trees of packed scalar stmts if SLP is possible.  */

bool
vect_analyze_slp (loop_vec_info loop_vinfo, bb_vec_info bb_vinfo)
{
  unsigned int i;
  vec<gimple> grouped_stores;
  vec<gimple> reductions = vNULL;
  vec<gimple> reduc_chains = vNULL;
  gimple first_element;
  bool ok = false;

  if (dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location, "=== vect_analyze_slp ===\n");

  if (loop_vinfo)
    {
      grouped_stores = LOOP_VINFO_GROUPED_STORES (loop_vinfo);
      reduc_chains = LOOP_VINFO_REDUCTION_CHAINS (loop_vinfo);
      reductions = LOOP_VINFO_REDUCTIONS (loop_vinfo);
    }
  else
    grouped_stores = BB_VINFO_GROUPED_STORES (bb_vinfo);

  /* Find SLP sequences starting from groups of grouped stores.  */
  FOR_EACH_VEC_ELT (grouped_stores, i, first_element)
    if (vect_analyze_slp_instance (loop_vinfo, bb_vinfo, first_element))
      ok = true;

  if (bb_vinfo && !ok)
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			 "Failed to SLP the basic block.\n");

      return false;
    }

  if (loop_vinfo
      && LOOP_VINFO_REDUCTION_CHAINS (loop_vinfo).length () > 0)
    {
      /* Find SLP sequences starting from reduction chains.  */
      FOR_EACH_VEC_ELT (reduc_chains, i, first_element)
        if (vect_analyze_slp_instance (loop_vinfo, bb_vinfo, first_element))
          ok = true;
        else
          return false;

      /* Don't try to vectorize SLP reductions if reduction chain was
         detected.  */
      return ok;
    }

  /* Find SLP sequences starting from groups of reductions.  */
  if (loop_vinfo && LOOP_VINFO_REDUCTIONS (loop_vinfo).length () > 1
      && vect_analyze_slp_instance (loop_vinfo, bb_vinfo, reductions[0]))
    ok = true;

  return true;
}


/* For each possible SLP instance decide whether to SLP it and calculate overall
   unrolling factor needed to SLP the loop.  Return TRUE if decided to SLP at
   least one instance.  */

bool
vect_make_slp_decision (loop_vec_info loop_vinfo)
{
  unsigned int i, unrolling_factor = 1;
  vec<slp_instance> slp_instances = LOOP_VINFO_SLP_INSTANCES (loop_vinfo);
  slp_instance instance;
  int decided_to_slp = 0;

  if (dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location, "=== vect_make_slp_decision ==="
                     "\n");

  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    {
      /* FORNOW: SLP if you can.  */
      if (unrolling_factor < SLP_INSTANCE_UNROLLING_FACTOR (instance))
	unrolling_factor = SLP_INSTANCE_UNROLLING_FACTOR (instance);

      /* Mark all the stmts that belong to INSTANCE as PURE_SLP stmts.  Later we
	 call vect_detect_hybrid_slp () to find stmts that need hybrid SLP and
	 loop-based vectorization.  Such stmts will be marked as HYBRID.  */
      vect_mark_slp_stmts (SLP_INSTANCE_TREE (instance), pure_slp, -1);
      decided_to_slp++;
    }

  LOOP_VINFO_SLP_UNROLLING_FACTOR (loop_vinfo) = unrolling_factor;

  if (decided_to_slp && dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location,
		     "Decided to SLP %d instances. Unrolling factor %d\n",
		     decided_to_slp, unrolling_factor);

  return (decided_to_slp > 0);
}


/* Find stmts that must be both vectorized and SLPed (since they feed stmts that
   can't be SLPed) in the tree rooted at NODE.  Mark such stmts as HYBRID.  */

static void
vect_detect_hybrid_slp_stmts (slp_tree node)
{
  int i;
  vec<gimple> stmts = SLP_TREE_SCALAR_STMTS (node);
  gimple stmt = stmts[0];
  imm_use_iterator imm_iter;
  gimple use_stmt;
  stmt_vec_info stmt_vinfo = vinfo_for_stmt (stmt);
  slp_tree child;
  loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_vinfo);
  struct loop *loop = NULL;
  bb_vec_info bb_vinfo = STMT_VINFO_BB_VINFO (stmt_vinfo);
  basic_block bb = NULL;

  if (!node)
    return;

  if (loop_vinfo)
    loop = LOOP_VINFO_LOOP (loop_vinfo);
  else
    bb = BB_VINFO_BB (bb_vinfo);

  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, stmt)
    if (PURE_SLP_STMT (vinfo_for_stmt (stmt))
	&& TREE_CODE (gimple_op (stmt, 0)) == SSA_NAME)
      FOR_EACH_IMM_USE_STMT (use_stmt, imm_iter, gimple_op (stmt, 0))
	if (gimple_bb (use_stmt)
            && ((loop && flow_bb_inside_loop_p (loop, gimple_bb (use_stmt)))
		 || bb == gimple_bb (use_stmt))
	    && (stmt_vinfo = vinfo_for_stmt (use_stmt))
	    && !STMT_SLP_TYPE (stmt_vinfo)
            && (STMT_VINFO_RELEVANT (stmt_vinfo)
                || VECTORIZABLE_CYCLE_DEF (STMT_VINFO_DEF_TYPE (stmt_vinfo)))
	    && !(gimple_code (use_stmt) == GIMPLE_PHI
                 && STMT_VINFO_DEF_TYPE (stmt_vinfo)
                  == vect_reduction_def))
	  vect_mark_slp_stmts (node, hybrid, i);

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_detect_hybrid_slp_stmts (child);
}


/* Find stmts that must be both vectorized and SLPed.  */

void
vect_detect_hybrid_slp (loop_vec_info loop_vinfo)
{
  unsigned int i;
  vec<slp_instance> slp_instances = LOOP_VINFO_SLP_INSTANCES (loop_vinfo);
  slp_instance instance;

  if (dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location, "=== vect_detect_hybrid_slp ==="
                     "\n");

  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    vect_detect_hybrid_slp_stmts (SLP_INSTANCE_TREE (instance));
}


/* Create and initialize a new bb_vec_info struct for BB, as well as
   stmt_vec_info structs for all the stmts in it.  */

static bb_vec_info
new_bb_vec_info (basic_block bb)
{
  bb_vec_info res = NULL;
  gimple_stmt_iterator gsi;

  res = (bb_vec_info) xcalloc (1, sizeof (struct _bb_vec_info));
  BB_VINFO_BB (res) = bb;

  for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (&gsi))
    {
      gimple stmt = gsi_stmt (gsi);
      gimple_set_uid (stmt, 0);
      set_vinfo_for_stmt (stmt, new_stmt_vec_info (stmt, NULL, res));
    }

  BB_VINFO_GROUPED_STORES (res).create (10);
  BB_VINFO_SLP_INSTANCES (res).create (2);
  BB_VINFO_TARGET_COST_DATA (res) = init_cost (NULL);

  bb->aux = res;
  return res;
}


/* Free BB_VINFO struct, as well as all the stmt_vec_info structs of all the
   stmts in the basic block.  */

static void
destroy_bb_vec_info (bb_vec_info bb_vinfo)
{
  vec<slp_instance> slp_instances;
  slp_instance instance;
  basic_block bb;
  gimple_stmt_iterator si;
  unsigned i;

  if (!bb_vinfo)
    return;

  bb = BB_VINFO_BB (bb_vinfo);

  for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
    {
      gimple stmt = gsi_stmt (si);
      stmt_vec_info stmt_info = vinfo_for_stmt (stmt);

      if (stmt_info)
        /* Free stmt_vec_info.  */
        free_stmt_vec_info (stmt);
    }

  vect_destroy_datarefs (NULL, bb_vinfo);
  free_dependence_relations (BB_VINFO_DDRS (bb_vinfo));
  BB_VINFO_GROUPED_STORES (bb_vinfo).release ();
  slp_instances = BB_VINFO_SLP_INSTANCES (bb_vinfo);
  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    vect_free_slp_instance (instance);
  BB_VINFO_SLP_INSTANCES (bb_vinfo).release ();
  destroy_cost_data (BB_VINFO_TARGET_COST_DATA (bb_vinfo));
  free (bb_vinfo);
  bb->aux = NULL;
}


/* Analyze statements contained in SLP tree node after recursively analyzing
   the subtree. Return TRUE if the operations are supported.  */

static bool
vect_slp_analyze_node_operations (bb_vec_info bb_vinfo, slp_tree node)
{
  bool dummy;
  int i;
  gimple stmt;
  slp_tree child;

  if (!node)
    return true;

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    if (!vect_slp_analyze_node_operations (bb_vinfo, child))
      return false;

  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, stmt)
    {
      stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
      gcc_assert (stmt_info);
      gcc_assert (PURE_SLP_STMT (stmt_info));

      if (!vect_analyze_stmt (stmt, &dummy, node))
        return false;
    }

  return true;
}


/* Analyze statements in SLP instances of the basic block.  Return TRUE if the
   operations are supported. */

static bool
vect_slp_analyze_operations (bb_vec_info bb_vinfo)
{
  vec<slp_instance> slp_instances = BB_VINFO_SLP_INSTANCES (bb_vinfo);
  slp_instance instance;
  int i;

  for (i = 0; slp_instances.iterate (i, &instance); )
    {
      if (!vect_slp_analyze_node_operations (bb_vinfo,
                                             SLP_INSTANCE_TREE (instance)))
        {
 	  vect_free_slp_instance (instance);
          slp_instances.ordered_remove (i);
	}
      else
        i++;
    }

  if (!slp_instances.length ())
    return false;

  return true;
}


/* Compute the scalar cost of the SLP node NODE and its children
   and return it.  Do not account defs that are marked in LIFE and
   update LIFE according to uses of NODE.  */

static unsigned
vect_bb_slp_scalar_cost (basic_block bb,
			 slp_tree node, vec<bool, va_heap> *life)
{
  unsigned scalar_cost = 0;
  unsigned i;
  gimple stmt;
  slp_tree child;

  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, stmt)
    {
      unsigned stmt_cost;
      ssa_op_iter op_iter;
      def_operand_p def_p;
      stmt_vec_info stmt_info;

      if ((*life)[i])
	continue;

      /* If there is a non-vectorized use of the defs then the scalar
         stmt is kept live in which case we do not account it or any
	 required defs in the SLP children in the scalar cost.  This
	 way we make the vectorization more costly when compared to
	 the scalar cost.  */
      FOR_EACH_SSA_DEF_OPERAND (def_p, stmt, op_iter, SSA_OP_DEF)
	{
	  imm_use_iterator use_iter;
	  gimple use_stmt;
	  FOR_EACH_IMM_USE_STMT (use_stmt, use_iter, DEF_FROM_PTR (def_p))
	    if (gimple_code (use_stmt) == GIMPLE_PHI
		|| gimple_bb (use_stmt) != bb
		|| !STMT_VINFO_VECTORIZABLE (vinfo_for_stmt (use_stmt)))
	      {
		(*life)[i] = true;
		BREAK_FROM_IMM_USE_STMT (use_iter);
	      }
	}
      if ((*life)[i])
	continue;

      stmt_info = vinfo_for_stmt (stmt);
      if (STMT_VINFO_DATA_REF (stmt_info))
        {
          if (DR_IS_READ (STMT_VINFO_DATA_REF (stmt_info)))
            stmt_cost = vect_get_stmt_cost (scalar_load);
          else
            stmt_cost = vect_get_stmt_cost (scalar_store);
        }
      else
        stmt_cost = vect_get_stmt_cost (scalar_stmt);

      scalar_cost += stmt_cost;
    }

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    scalar_cost += vect_bb_slp_scalar_cost (bb, child, life);

  return scalar_cost;
}

/* Check if vectorization of the basic block is profitable.  */

static bool
vect_bb_vectorization_profitable_p (bb_vec_info bb_vinfo)
{
  vec<slp_instance> slp_instances = BB_VINFO_SLP_INSTANCES (bb_vinfo);
  slp_instance instance;
  int i, j;
  unsigned int vec_inside_cost = 0, vec_outside_cost = 0, scalar_cost = 0;
  unsigned int vec_prologue_cost = 0, vec_epilogue_cost = 0;
  void *target_cost_data = BB_VINFO_TARGET_COST_DATA (bb_vinfo);
  stmt_vec_info stmt_info = NULL;
  stmt_vector_for_cost body_cost_vec;
  stmt_info_for_cost *ci;

  /* Calculate vector costs.  */
  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    {
      body_cost_vec = SLP_INSTANCE_BODY_COST_VEC (instance);

      FOR_EACH_VEC_ELT (body_cost_vec, j, ci)
	{
	  stmt_info = ci->stmt ? vinfo_for_stmt (ci->stmt) : NULL;
	  (void) add_stmt_cost (target_cost_data, ci->count, ci->kind,
				stmt_info, ci->misalign, vect_body);
	}
    }

  /* Calculate scalar cost.  */
  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    {
      auto_vec<bool, 20> life;
      life.safe_grow_cleared (SLP_INSTANCE_GROUP_SIZE (instance));
      scalar_cost += vect_bb_slp_scalar_cost (BB_VINFO_BB (bb_vinfo),
					      SLP_INSTANCE_TREE (instance),
					      &life);
    }

  /* Complete the target-specific cost calculation.  */
  finish_cost (BB_VINFO_TARGET_COST_DATA (bb_vinfo), &vec_prologue_cost,
	       &vec_inside_cost, &vec_epilogue_cost);

  vec_outside_cost = vec_prologue_cost + vec_epilogue_cost;

  if (dump_enabled_p ())
    {
      dump_printf_loc (MSG_NOTE, vect_location, "Cost model analysis: \n");
      dump_printf (MSG_NOTE, "  Vector inside of basic block cost: %d\n",
		   vec_inside_cost);
      dump_printf (MSG_NOTE, "  Vector prologue cost: %d\n", vec_prologue_cost);
      dump_printf (MSG_NOTE, "  Vector epilogue cost: %d\n", vec_epilogue_cost);
      dump_printf (MSG_NOTE, "  Scalar cost of basic block: %d\n", scalar_cost);
    }

  /* Vectorization is profitable if its cost is less than the cost of scalar
     version.  */
  if (vec_outside_cost + vec_inside_cost >= scalar_cost)
    return false;

  return true;
}

/* Check if the basic block can be vectorized.  */

static bb_vec_info
vect_slp_analyze_bb_1 (basic_block bb)
{
  bb_vec_info bb_vinfo;
  vec<slp_instance> slp_instances;
  slp_instance instance;
  int i;
  int min_vf = 2;

  bb_vinfo = new_bb_vec_info (bb);
  if (!bb_vinfo)
    return NULL;

  if (!vect_analyze_data_refs (NULL, bb_vinfo, &min_vf))
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			 "not vectorized: unhandled data-ref in basic "
			 "block.\n");

      destroy_bb_vec_info (bb_vinfo);
      return NULL;
    }

  if (BB_VINFO_DATAREFS (bb_vinfo).length () < 2)
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			 "not vectorized: not enough data-refs in "
			 "basic block.\n");

      destroy_bb_vec_info (bb_vinfo);
      return NULL;
    }

  if (!vect_analyze_data_ref_accesses (NULL, bb_vinfo))
    {
     if (dump_enabled_p ())
       dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			"not vectorized: unhandled data access in "
			"basic block.\n");

      destroy_bb_vec_info (bb_vinfo);
      return NULL;
    }

  vect_pattern_recog (NULL, bb_vinfo);

  if (!vect_slp_analyze_data_ref_dependences (bb_vinfo))
     {
       if (dump_enabled_p ())
	 dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			  "not vectorized: unhandled data dependence "
			  "in basic block.\n");

       destroy_bb_vec_info (bb_vinfo);
       return NULL;
     }

  if (!vect_analyze_data_refs_alignment (NULL, bb_vinfo))
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			 "not vectorized: bad data alignment in basic "
			 "block.\n");

      destroy_bb_vec_info (bb_vinfo);
      return NULL;
    }

  /* Check the SLP opportunities in the basic block, analyze and build SLP
     trees.  */
  if (!vect_analyze_slp (NULL, bb_vinfo))
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location, 
			 "not vectorized: failed to find SLP opportunities "
			 "in basic block.\n");

      destroy_bb_vec_info (bb_vinfo);
      return NULL;
    }

  slp_instances = BB_VINFO_SLP_INSTANCES (bb_vinfo);

  /* Mark all the statements that we want to vectorize as pure SLP and
     relevant.  */
  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    {
      vect_mark_slp_stmts (SLP_INSTANCE_TREE (instance), pure_slp, -1);
      vect_mark_slp_stmts_relevant (SLP_INSTANCE_TREE (instance));
    }

  if (!vect_verify_datarefs_alignment (NULL, bb_vinfo))
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
                         "not vectorized: unsupported alignment in basic "
                         "block.\n");
      destroy_bb_vec_info (bb_vinfo);
      return NULL;
    }

  if (!vect_slp_analyze_operations (bb_vinfo))
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			 "not vectorized: bad operation in basic block.\n");

      destroy_bb_vec_info (bb_vinfo);
      return NULL;
    }

  /* Cost model: check if the vectorization is worthwhile.  */
  if (!unlimited_cost_model (NULL)
      && !vect_bb_vectorization_profitable_p (bb_vinfo))
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			 "not vectorized: vectorization is not "
			 "profitable.\n");

      destroy_bb_vec_info (bb_vinfo);
      return NULL;
    }

  if (dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location,
		     "Basic block will be vectorized using SLP\n");

  return bb_vinfo;
}


bb_vec_info
vect_slp_analyze_bb (basic_block bb)
{
  bb_vec_info bb_vinfo;
  int insns = 0;
  gimple_stmt_iterator gsi;
  unsigned int vector_sizes;

  if (dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location, "===vect_slp_analyze_bb===\n");

  for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (&gsi))
    {
      gimple stmt = gsi_stmt (gsi);
      if (!is_gimple_debug (stmt)
          && !gimple_nop_p (stmt)
          && gimple_code (stmt) != GIMPLE_LABEL)
        insns++;
    }

  if (insns > PARAM_VALUE (PARAM_SLP_MAX_INSNS_IN_BB))
    {
      if (dump_enabled_p ())
        dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			 "not vectorized: too many instructions in "
			 "basic block.\n");

      return NULL;
    }

  /* Autodetect first vector size we try.  */
  current_vector_size = 0;
  vector_sizes = targetm.vectorize.autovectorize_vector_sizes ();

  while (1)
    {
      bb_vinfo = vect_slp_analyze_bb_1 (bb);
      if (bb_vinfo)
        return bb_vinfo;

      destroy_bb_vec_info (bb_vinfo);

      vector_sizes &= ~current_vector_size;
      if (vector_sizes == 0
          || current_vector_size == 0)
        return NULL;

      /* Try the next biggest vector size.  */
      current_vector_size = 1 << floor_log2 (vector_sizes);
      if (dump_enabled_p ())
        dump_printf_loc (MSG_NOTE, vect_location,
			 "***** Re-trying analysis with "
			 "vector size %d\n", current_vector_size);
    }
}


/* SLP costs are calculated according to SLP instance unrolling factor (i.e.,
   the number of created vector stmts depends on the unrolling factor).
   However, the actual number of vector stmts for every SLP node depends on
   VF which is set later in vect_analyze_operations ().  Hence, SLP costs
   should be updated.  In this function we assume that the inside costs
   calculated in vect_model_xxx_cost are linear in ncopies.  */

void
vect_update_slp_costs_according_to_vf (loop_vec_info loop_vinfo)
{
  unsigned int i, j, vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
  vec<slp_instance> slp_instances = LOOP_VINFO_SLP_INSTANCES (loop_vinfo);
  slp_instance instance;
  stmt_vector_for_cost body_cost_vec;
  stmt_info_for_cost *si;
  void *data = LOOP_VINFO_TARGET_COST_DATA (loop_vinfo);

  if (dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location,
		     "=== vect_update_slp_costs_according_to_vf ===\n");

  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    {
      /* We assume that costs are linear in ncopies.  */
      int ncopies = vf / SLP_INSTANCE_UNROLLING_FACTOR (instance);

      /* Record the instance's instructions in the target cost model.
	 This was delayed until here because the count of instructions
	 isn't known beforehand.  */
      body_cost_vec = SLP_INSTANCE_BODY_COST_VEC (instance);

      FOR_EACH_VEC_ELT (body_cost_vec, j, si)
	(void) add_stmt_cost (data, si->count * ncopies, si->kind,
			      vinfo_for_stmt (si->stmt), si->misalign,
			      vect_body);
    }
}


/* For constant and loop invariant defs of SLP_NODE this function returns
   (vector) defs (VEC_OPRNDS) that will be used in the vectorized stmts.
   OP_NUM determines if we gather defs for operand 0 or operand 1 of the RHS of
   scalar stmts.  NUMBER_OF_VECTORS is the number of vector defs to create.
   REDUC_INDEX is the index of the reduction operand in the statements, unless
   it is -1.  */

static void
vect_get_constant_vectors (tree op, slp_tree slp_node,
                           vec<tree> *vec_oprnds,
			   unsigned int op_num, unsigned int number_of_vectors,
                           int reduc_index)
{
  vec<gimple> stmts = SLP_TREE_SCALAR_STMTS (slp_node);
  gimple stmt = stmts[0];
  stmt_vec_info stmt_vinfo = vinfo_for_stmt (stmt);
  unsigned nunits;
  tree vec_cst;
  tree *elts;
  unsigned j, number_of_places_left_in_vector;
  tree vector_type;
  tree vop;
  int group_size = stmts.length ();
  unsigned int vec_num, i;
  unsigned number_of_copies = 1;
  vec<tree> voprnds;
  voprnds.create (number_of_vectors);
  bool constant_p, is_store;
  tree neutral_op = NULL;
  enum tree_code code = gimple_expr_code (stmt);
  gimple def_stmt;
  struct loop *loop;
  gimple_seq ctor_seq = NULL;

  if (STMT_VINFO_DEF_TYPE (stmt_vinfo) == vect_reduction_def
      && reduc_index != -1)
    {
      op_num = reduc_index - 1;
      op = gimple_op (stmt, reduc_index);
      /* For additional copies (see the explanation of NUMBER_OF_COPIES below)
         we need either neutral operands or the original operands.  See
         get_initial_def_for_reduction() for details.  */
      switch (code)
        {
          case WIDEN_SUM_EXPR:
          case DOT_PROD_EXPR:
          case PLUS_EXPR:
          case MINUS_EXPR:
          case BIT_IOR_EXPR:
          case BIT_XOR_EXPR:
             if (SCALAR_FLOAT_TYPE_P (TREE_TYPE (op)))
               neutral_op = build_real (TREE_TYPE (op), dconst0);
             else
               neutral_op = build_int_cst (TREE_TYPE (op), 0);

             break;

          case MULT_EXPR:
             if (SCALAR_FLOAT_TYPE_P (TREE_TYPE (op)))
               neutral_op = build_real (TREE_TYPE (op), dconst1);
             else
               neutral_op = build_int_cst (TREE_TYPE (op), 1);

             break;

          case BIT_AND_EXPR:
            neutral_op = build_int_cst (TREE_TYPE (op), -1);
            break;

          case MAX_EXPR:
          case MIN_EXPR:
            def_stmt = SSA_NAME_DEF_STMT (op);
            loop = (gimple_bb (stmt))->loop_father;
            neutral_op = PHI_ARG_DEF_FROM_EDGE (def_stmt,
                                                loop_preheader_edge (loop));
            break;

          default:
            neutral_op = NULL;
        }
    }

  if (STMT_VINFO_DATA_REF (stmt_vinfo))
    {
      is_store = true;
      op = gimple_assign_rhs1 (stmt);
    }
  else
    is_store = false;

  gcc_assert (op);

  if (CONSTANT_CLASS_P (op))
    constant_p = true;
  else
    constant_p = false;

  vector_type = get_vectype_for_scalar_type (TREE_TYPE (op));
  gcc_assert (vector_type);
  nunits = TYPE_VECTOR_SUBPARTS (vector_type);

  /* NUMBER_OF_COPIES is the number of times we need to use the same values in
     created vectors. It is greater than 1 if unrolling is performed.

     For example, we have two scalar operands, s1 and s2 (e.g., group of
     strided accesses of size two), while NUNITS is four (i.e., four scalars
     of this type can be packed in a vector).  The output vector will contain
     two copies of each scalar operand: {s1, s2, s1, s2}.  (NUMBER_OF_COPIES
     will be 2).

     If GROUP_SIZE > NUNITS, the scalars will be split into several vectors
     containing the operands.

     For example, NUNITS is four as before, and the group size is 8
     (s1, s2, ..., s8).  We will create two vectors {s1, s2, s3, s4} and
     {s5, s6, s7, s8}.  */

  number_of_copies = least_common_multiple (nunits, group_size) / group_size;

  number_of_places_left_in_vector = nunits;
  elts = XALLOCAVEC (tree, nunits);
  for (j = 0; j < number_of_copies; j++)
    {
      for (i = group_size - 1; stmts.iterate (i, &stmt); i--)
        {
          if (is_store)
            op = gimple_assign_rhs1 (stmt);
          else
	    {
	      switch (code)
		{
		  case COND_EXPR:
		    if (op_num == 0 || op_num == 1)
		      {
			tree cond = gimple_assign_rhs1 (stmt);
			op = TREE_OPERAND (cond, op_num);
		      }
		    else
		      {
			if (op_num == 2)
			  op = gimple_assign_rhs2 (stmt);
			else
			  op = gimple_assign_rhs3 (stmt);
		      }
		    break;

		  case CALL_EXPR:
		    op = gimple_call_arg (stmt, op_num);
		    break;

		  case LSHIFT_EXPR:
		  case RSHIFT_EXPR:
		  case LROTATE_EXPR:
		  case RROTATE_EXPR:
		    op = gimple_op (stmt, op_num + 1);
		    /* Unlike the other binary operators, shifts/rotates have
		       the shift count being int, instead of the same type as
		       the lhs, so make sure the scalar is the right type if
		       we are dealing with vectors of
		       long long/long/short/char.  */
		    if (op_num == 1 && TREE_CODE (op) == INTEGER_CST)
		      op = fold_convert (TREE_TYPE (vector_type), op);
		    break;

		  default:
		    op = gimple_op (stmt, op_num + 1);
		    break;
		}
	    }

          if (reduc_index != -1)
            {
              loop = (gimple_bb (stmt))->loop_father;
              def_stmt = SSA_NAME_DEF_STMT (op);

              gcc_assert (loop);

              /* Get the def before the loop.  In reduction chain we have only
                 one initial value.  */
              if ((j != (number_of_copies - 1)
                   || (GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt))
                       && i != 0))
                  && neutral_op)
                op = neutral_op;
              else
                op = PHI_ARG_DEF_FROM_EDGE (def_stmt,
                                            loop_preheader_edge (loop));
            }

          /* Create 'vect_ = {op0,op1,...,opn}'.  */
          number_of_places_left_in_vector--;
	  if (!types_compatible_p (TREE_TYPE (vector_type), TREE_TYPE (op)))
	    {
	      if (CONSTANT_CLASS_P (op))
		{
		  op = fold_unary (VIEW_CONVERT_EXPR,
				   TREE_TYPE (vector_type), op);
		  gcc_assert (op && CONSTANT_CLASS_P (op));
		}
	      else
		{
		  tree new_temp
		    = make_ssa_name (TREE_TYPE (vector_type), NULL);
		  gimple init_stmt;
		  op = build1 (VIEW_CONVERT_EXPR, TREE_TYPE (vector_type),
			       op);		  
		  init_stmt
		    = gimple_build_assign_with_ops (VIEW_CONVERT_EXPR,
						    new_temp, op, NULL_TREE);
		  gimple_seq_add_stmt (&ctor_seq, init_stmt);
		  op = new_temp;
		}
	    }
	  elts[number_of_places_left_in_vector] = op;
	  if (!CONSTANT_CLASS_P (op))
	    constant_p = false;

          if (number_of_places_left_in_vector == 0)
            {
              number_of_places_left_in_vector = nunits;

	      if (constant_p)
		vec_cst = build_vector (vector_type, elts);
	      else
		{
		  vec<constructor_elt, va_gc> *v;
		  unsigned k;
		  vec_alloc (v, nunits);
		  for (k = 0; k < nunits; ++k)
		    CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, elts[k]);
		  vec_cst = build_constructor (vector_type, v);
		}
              voprnds.quick_push (vect_init_vector (stmt, vec_cst,
						    vector_type, NULL));
	      if (ctor_seq != NULL)
		{
		  gimple init_stmt = SSA_NAME_DEF_STMT (voprnds.last ());
		  gimple_stmt_iterator gsi = gsi_for_stmt (init_stmt);
		  gsi_insert_seq_before_without_update (&gsi, ctor_seq,
							GSI_SAME_STMT);
		  ctor_seq = NULL;
		}
            }
        }
    }

  /* Since the vectors are created in the reverse order, we should invert
     them.  */
  vec_num = voprnds.length ();
  for (j = vec_num; j != 0; j--)
    {
      vop = voprnds[j - 1];
      vec_oprnds->quick_push (vop);
    }

  voprnds.release ();

  /* In case that VF is greater than the unrolling factor needed for the SLP
     group of stmts, NUMBER_OF_VECTORS to be created is greater than
     NUMBER_OF_SCALARS/NUNITS or NUNITS/NUMBER_OF_SCALARS, and hence we have
     to replicate the vectors.  */
  while (number_of_vectors > vec_oprnds->length ())
    {
      tree neutral_vec = NULL;

      if (neutral_op)
        {
          if (!neutral_vec)
	    neutral_vec = build_vector_from_val (vector_type, neutral_op);

          vec_oprnds->quick_push (neutral_vec);
        }
      else
        {
          for (i = 0; vec_oprnds->iterate (i, &vop) && i < vec_num; i++)
            vec_oprnds->quick_push (vop);
        }
    }
}


/* Get vectorized definitions from SLP_NODE that contains corresponding
   vectorized def-stmts.  */

static void
vect_get_slp_vect_defs (slp_tree slp_node, vec<tree> *vec_oprnds)
{
  tree vec_oprnd;
  gimple vec_def_stmt;
  unsigned int i;

  gcc_assert (SLP_TREE_VEC_STMTS (slp_node).exists ());

  FOR_EACH_VEC_ELT (SLP_TREE_VEC_STMTS (slp_node), i, vec_def_stmt)
    {
      gcc_assert (vec_def_stmt);
      vec_oprnd = gimple_get_lhs (vec_def_stmt);
      vec_oprnds->quick_push (vec_oprnd);
    }
}


/* Get vectorized definitions for SLP_NODE.
   If the scalar definitions are loop invariants or constants, collect them and
   call vect_get_constant_vectors() to create vector stmts.
   Otherwise, the def-stmts must be already vectorized and the vectorized stmts
   must be stored in the corresponding child of SLP_NODE, and we call
   vect_get_slp_vect_defs () to retrieve them.  */

void
vect_get_slp_defs (vec<tree> ops, slp_tree slp_node,
		   vec<vec<tree> > *vec_oprnds, int reduc_index)
{
  gimple first_stmt;
  int number_of_vects = 0, i;
  unsigned int child_index = 0;
  HOST_WIDE_INT lhs_size_unit, rhs_size_unit;
  slp_tree child = NULL;
  vec<tree> vec_defs;
  tree oprnd;
  bool vectorized_defs;

  first_stmt = SLP_TREE_SCALAR_STMTS (slp_node)[0];
  FOR_EACH_VEC_ELT (ops, i, oprnd)
    {
      /* For each operand we check if it has vectorized definitions in a child
	 node or we need to create them (for invariants and constants).  We
	 check if the LHS of the first stmt of the next child matches OPRND.
	 If it does, we found the correct child.  Otherwise, we call
	 vect_get_constant_vectors (), and not advance CHILD_INDEX in order
	 to check this child node for the next operand.  */
      vectorized_defs = false;
      if (SLP_TREE_CHILDREN (slp_node).length () > child_index)
        {
          child = SLP_TREE_CHILDREN (slp_node)[child_index];

	  /* We have to check both pattern and original def, if available.  */
	  gimple first_def = SLP_TREE_SCALAR_STMTS (child)[0];
	  gimple related = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (first_def));

	  if (operand_equal_p (oprnd, gimple_get_lhs (first_def), 0)
	      || (related
		  && operand_equal_p (oprnd, gimple_get_lhs (related), 0)))
	    {
	      /* The number of vector defs is determined by the number of
		 vector statements in the node from which we get those
		 statements.  */
	      number_of_vects = SLP_TREE_NUMBER_OF_VEC_STMTS (child);
	      vectorized_defs = true;
	      child_index++;
	    }
        }

      if (!vectorized_defs)
        {
          if (i == 0)
            {
              number_of_vects = SLP_TREE_NUMBER_OF_VEC_STMTS (slp_node);
              /* Number of vector stmts was calculated according to LHS in
                 vect_schedule_slp_instance (), fix it by replacing LHS with
                 RHS, if necessary.  See vect_get_smallest_scalar_type () for
                 details.  */
              vect_get_smallest_scalar_type (first_stmt, &lhs_size_unit,
                                             &rhs_size_unit);
              if (rhs_size_unit != lhs_size_unit)
                {
                  number_of_vects *= rhs_size_unit;
                  number_of_vects /= lhs_size_unit;
                }
            }
        }

      /* Allocate memory for vectorized defs.  */
      vec_defs = vNULL;
      vec_defs.create (number_of_vects);

      /* For reduction defs we call vect_get_constant_vectors (), since we are
         looking for initial loop invariant values.  */
      if (vectorized_defs && reduc_index == -1)
        /* The defs are already vectorized.  */
	vect_get_slp_vect_defs (child, &vec_defs);
      else
        /* Build vectors from scalar defs.  */
	vect_get_constant_vectors (oprnd, slp_node, &vec_defs, i,
                                   number_of_vects, reduc_index);

      vec_oprnds->quick_push (vec_defs);

      /* For reductions, we only need initial values.  */
      if (reduc_index != -1)
        return;
    }
}


/* Create NCOPIES permutation statements using the mask MASK_BYTES (by
   building a vector of type MASK_TYPE from it) and two input vectors placed in
   DR_CHAIN at FIRST_VEC_INDX and SECOND_VEC_INDX for the first copy and
   shifting by STRIDE elements of DR_CHAIN for every copy.
   (STRIDE is the number of vectorized stmts for NODE divided by the number of
   copies).
   VECT_STMTS_COUNTER specifies the index in the vectorized stmts of NODE, where
   the created stmts must be inserted.  */

static inline void
vect_create_mask_and_perm (gimple stmt, gimple next_scalar_stmt,
                           tree mask, int first_vec_indx, int second_vec_indx,
                           gimple_stmt_iterator *gsi, slp_tree node,
                           tree vectype, vec<tree> dr_chain,
                           int ncopies, int vect_stmts_counter)
{
  tree perm_dest;
  gimple perm_stmt = NULL;
  stmt_vec_info next_stmt_info;
  int i, stride;
  tree first_vec, second_vec, data_ref;

  stride = SLP_TREE_NUMBER_OF_VEC_STMTS (node) / ncopies;

  /* Initialize the vect stmts of NODE to properly insert the generated
     stmts later.  */
  for (i = SLP_TREE_VEC_STMTS (node).length ();
       i < (int) SLP_TREE_NUMBER_OF_VEC_STMTS (node); i++)
    SLP_TREE_VEC_STMTS (node).quick_push (NULL);

  perm_dest = vect_create_destination_var (gimple_assign_lhs (stmt), vectype);
  for (i = 0; i < ncopies; i++)
    {
      first_vec = dr_chain[first_vec_indx];
      second_vec = dr_chain[second_vec_indx];

      /* Generate the permute statement.  */
      perm_stmt = gimple_build_assign_with_ops (VEC_PERM_EXPR, perm_dest,
						first_vec, second_vec, mask);
      data_ref = make_ssa_name (perm_dest, perm_stmt);
      gimple_set_lhs (perm_stmt, data_ref);
      vect_finish_stmt_generation (stmt, perm_stmt, gsi);

      /* Store the vector statement in NODE.  */
      SLP_TREE_VEC_STMTS (node)[stride * i + vect_stmts_counter] = perm_stmt;

      first_vec_indx += stride;
      second_vec_indx += stride;
    }

  /* Mark the scalar stmt as vectorized.  */
  next_stmt_info = vinfo_for_stmt (next_scalar_stmt);
  STMT_VINFO_VEC_STMT (next_stmt_info) = perm_stmt;
}


/* Given FIRST_MASK_ELEMENT - the mask element in element representation,
   return in CURRENT_MASK_ELEMENT its equivalent in target specific
   representation.  Check that the mask is valid and return FALSE if not.
   Return TRUE in NEED_NEXT_VECTOR if the permutation requires to move to
   the next vector, i.e., the current first vector is not needed.  */

static bool
vect_get_mask_element (gimple stmt, int first_mask_element, int m,
                       int mask_nunits, bool only_one_vec, int index,
		       unsigned char *mask, int *current_mask_element,
                       bool *need_next_vector, int *number_of_mask_fixes,
                       bool *mask_fixed, bool *needs_first_vector)
{
  int i;

  /* Convert to target specific representation.  */
  *current_mask_element = first_mask_element + m;
  /* Adjust the value in case it's a mask for second and third vectors.  */
  *current_mask_element -= mask_nunits * (*number_of_mask_fixes - 1);

  if (*current_mask_element < mask_nunits)
    *needs_first_vector = true;

  /* We have only one input vector to permute but the mask accesses values in
     the next vector as well.  */
  if (only_one_vec && *current_mask_element >= mask_nunits)
    {
      if (dump_enabled_p ())
        {
          dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			   "permutation requires at least two vectors ");
          dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
          dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
        }

      return false;
    }

  /* The mask requires the next vector.  */
  if (*current_mask_element >= mask_nunits * 2)
    {
      if (*needs_first_vector || *mask_fixed)
        {
          /* We either need the first vector too or have already moved to the
             next vector. In both cases, this permutation needs three
             vectors.  */
          if (dump_enabled_p ())
            {
              dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			       "permutation requires at "
			       "least three vectors ");
              dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
              dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
            }

          return false;
        }

      /* We move to the next vector, dropping the first one and working with
         the second and the third - we need to adjust the values of the mask
         accordingly.  */
      *current_mask_element -= mask_nunits * *number_of_mask_fixes;

      for (i = 0; i < index; i++)
        mask[i] -= mask_nunits * *number_of_mask_fixes;

      (*number_of_mask_fixes)++;
      *mask_fixed = true;
    }

  *need_next_vector = *mask_fixed;

  /* This was the last element of this mask. Start a new one.  */
  if (index == mask_nunits - 1)
    {
      *number_of_mask_fixes = 1;
      *mask_fixed = false;
      *needs_first_vector = false;
    }

  return true;
}


/* Generate vector permute statements from a list of loads in DR_CHAIN.
   If ANALYZE_ONLY is TRUE, only check that it is possible to create valid
   permute statements for the SLP node NODE of the SLP instance
   SLP_NODE_INSTANCE.  */

bool
vect_transform_slp_perm_load (slp_tree node, vec<tree> dr_chain,
                              gimple_stmt_iterator *gsi, int vf,
                              slp_instance slp_node_instance, bool analyze_only)
{
  gimple stmt = SLP_TREE_SCALAR_STMTS (node)[0];
  stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
  tree mask_element_type = NULL_TREE, mask_type;
  int i, j, k, nunits, vec_index = 0, scalar_index;
  tree vectype = STMT_VINFO_VECTYPE (stmt_info);
  gimple next_scalar_stmt;
  int group_size = SLP_INSTANCE_GROUP_SIZE (slp_node_instance);
  int first_mask_element;
  int index, unroll_factor, current_mask_element, ncopies;
  unsigned char *mask;
  bool only_one_vec = false, need_next_vector = false;
  int first_vec_index, second_vec_index, orig_vec_stmts_num, vect_stmts_counter;
  int number_of_mask_fixes = 1;
  bool mask_fixed = false;
  bool needs_first_vector = false;
  enum machine_mode mode;

  mode = TYPE_MODE (vectype);

  if (!can_vec_perm_p (mode, false, NULL))
    {
      if (dump_enabled_p ())
        {
          dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
			   "no vect permute for ");
          dump_gimple_stmt (MSG_MISSED_OPTIMIZATION, TDF_SLIM, stmt, 0);
          dump_printf (MSG_MISSED_OPTIMIZATION, "\n");
        }
      return false;
    }

  /* The generic VEC_PERM_EXPR code always uses an integral type of the
     same size as the vector element being permuted.  */
  mask_element_type = lang_hooks.types.type_for_mode
		(int_mode_for_mode (TYPE_MODE (TREE_TYPE (vectype))), 1);
  mask_type = get_vectype_for_scalar_type (mask_element_type);
  nunits = TYPE_VECTOR_SUBPARTS (vectype);
  mask = XALLOCAVEC (unsigned char, nunits);
  unroll_factor = SLP_INSTANCE_UNROLLING_FACTOR (slp_node_instance);

  /* The number of vector stmts to generate based only on SLP_NODE_INSTANCE
     unrolling factor.  */
  orig_vec_stmts_num = group_size *
                SLP_INSTANCE_UNROLLING_FACTOR (slp_node_instance) / nunits;
  if (orig_vec_stmts_num == 1)
    only_one_vec = true;

  /* Number of copies is determined by the final vectorization factor
     relatively to SLP_NODE_INSTANCE unrolling factor.  */
  ncopies = vf / SLP_INSTANCE_UNROLLING_FACTOR (slp_node_instance);

  if (!STMT_VINFO_GROUPED_ACCESS (stmt_info))
    return false;

  /* Generate permutation masks for every NODE. Number of masks for each NODE
     is equal to GROUP_SIZE.
     E.g., we have a group of three nodes with three loads from the same
     location in each node, and the vector size is 4. I.e., we have a
     a0b0c0a1b1c1... sequence and we need to create the following vectors:
     for a's: a0a0a0a1 a1a1a2a2 a2a3a3a3
     for b's: b0b0b0b1 b1b1b2b2 b2b3b3b3
     ...

     The masks for a's should be: {0,0,0,3} {3,3,6,6} {6,9,9,9}.
     The last mask is illegal since we assume two operands for permute
     operation, and the mask element values can't be outside that range.
     Hence, the last mask must be converted into {2,5,5,5}.
     For the first two permutations we need the first and the second input
     vectors: {a0,b0,c0,a1} and {b1,c1,a2,b2}, and for the last permutation
     we need the second and the third vectors: {b1,c1,a2,b2} and
     {c2,a3,b3,c3}.  */

    {
      scalar_index = 0;
      index = 0;
      vect_stmts_counter = 0;
      vec_index = 0;
      first_vec_index = vec_index++;
      if (only_one_vec)
        second_vec_index = first_vec_index;
      else
        second_vec_index =  vec_index++;

      for (j = 0; j < unroll_factor; j++)
        {
          for (k = 0; k < group_size; k++)
            {
	      i = SLP_TREE_LOAD_PERMUTATION (node)[k];
              first_mask_element = i + j * group_size;
              if (!vect_get_mask_element (stmt, first_mask_element, 0,
					  nunits, only_one_vec, index,
					  mask, &current_mask_element,
					  &need_next_vector,
					  &number_of_mask_fixes, &mask_fixed,
					  &needs_first_vector))
		return false;
	      mask[index++] = current_mask_element;

              if (index == nunits)
                {
		  index = 0;
		  if (!can_vec_perm_p (mode, false, mask))
		    {
		      if (dump_enabled_p ())
			{
			  dump_printf_loc (MSG_MISSED_OPTIMIZATION,
					   vect_location, 
					   "unsupported vect permute { ");
			  for (i = 0; i < nunits; ++i)
			    dump_printf (MSG_MISSED_OPTIMIZATION, "%d ",
					 mask[i]);
			  dump_printf (MSG_MISSED_OPTIMIZATION, "}\n");
			}
		      return false;
		    }

                  if (!analyze_only)
                    {
		      int l;
		      tree mask_vec, *mask_elts;
		      mask_elts = XALLOCAVEC (tree, nunits);
		      for (l = 0; l < nunits; ++l)
			mask_elts[l] = build_int_cst (mask_element_type,
						      mask[l]);
		      mask_vec = build_vector (mask_type, mask_elts);

		      if (need_next_vector)
                        {
                          first_vec_index = second_vec_index;
                          second_vec_index = vec_index;
                        }

                      next_scalar_stmt
			  = SLP_TREE_SCALAR_STMTS (node)[scalar_index++];

                      vect_create_mask_and_perm (stmt, next_scalar_stmt,
                               mask_vec, first_vec_index, second_vec_index,
			       gsi, node, vectype, dr_chain,
			       ncopies, vect_stmts_counter++);
                    }
                }
            }
        }
    }

  return true;
}



/* Vectorize SLP instance tree in postorder.  */

static bool
vect_schedule_slp_instance (slp_tree node, slp_instance instance,
                            unsigned int vectorization_factor)
{
  gimple stmt;
  bool grouped_store, is_store;
  gimple_stmt_iterator si;
  stmt_vec_info stmt_info;
  unsigned int vec_stmts_size, nunits, group_size;
  tree vectype;
  int i;
  slp_tree child;

  if (!node)
    return false;

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_schedule_slp_instance (child, instance, vectorization_factor);

  stmt = SLP_TREE_SCALAR_STMTS (node)[0];
  stmt_info = vinfo_for_stmt (stmt);

  /* VECTYPE is the type of the destination.  */
  vectype = STMT_VINFO_VECTYPE (stmt_info);
  nunits = (unsigned int) TYPE_VECTOR_SUBPARTS (vectype);
  group_size = SLP_INSTANCE_GROUP_SIZE (instance);

  /* For each SLP instance calculate number of vector stmts to be created
     for the scalar stmts in each node of the SLP tree.  Number of vector
     elements in one vector iteration is the number of scalar elements in
     one scalar iteration (GROUP_SIZE) multiplied by VF divided by vector
     size.  */
  vec_stmts_size = (vectorization_factor * group_size) / nunits;

  if (!SLP_TREE_VEC_STMTS (node).exists ())
    {
      SLP_TREE_VEC_STMTS (node).create (vec_stmts_size);
      SLP_TREE_NUMBER_OF_VEC_STMTS (node) = vec_stmts_size;
    }

  if (dump_enabled_p ())
    {
      dump_printf_loc (MSG_NOTE,vect_location,
		       "------>vectorizing SLP node starting from: ");
      dump_gimple_stmt (MSG_NOTE, TDF_SLIM, stmt, 0);
      dump_printf (MSG_NOTE, "\n");
    }

  /* Loads should be inserted before the first load.  */
  if (SLP_INSTANCE_FIRST_LOAD_STMT (instance)
      && STMT_VINFO_GROUPED_ACCESS (stmt_info)
      && !REFERENCE_CLASS_P (gimple_get_lhs (stmt))
      && SLP_TREE_LOAD_PERMUTATION (node).exists ())
    si = gsi_for_stmt (SLP_INSTANCE_FIRST_LOAD_STMT (instance));
  else if (is_pattern_stmt_p (stmt_info))
    si = gsi_for_stmt (STMT_VINFO_RELATED_STMT (stmt_info));
  else
    si = gsi_for_stmt (stmt);

  /* Stores should be inserted just before the last store.  */
  if (STMT_VINFO_GROUPED_ACCESS (stmt_info)
      && REFERENCE_CLASS_P (gimple_get_lhs (stmt)))
    { 
      gimple last_store = vect_find_last_store_in_slp_instance (instance);
      if (is_pattern_stmt_p (vinfo_for_stmt (last_store)))
       last_store = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (last_store));
      si = gsi_for_stmt (last_store);
    }

  /* Mark the first element of the reduction chain as reduction to properly
     transform the node.  In the analysis phase only the last element of the
     chain is marked as reduction.  */
  if (GROUP_FIRST_ELEMENT (stmt_info) && !STMT_VINFO_GROUPED_ACCESS (stmt_info)
      && GROUP_FIRST_ELEMENT (stmt_info) == stmt)
    {
      STMT_VINFO_DEF_TYPE (stmt_info) = vect_reduction_def;
      STMT_VINFO_TYPE (stmt_info) = reduc_vec_info_type;
    }

  is_store = vect_transform_stmt (stmt, &si, &grouped_store, node, instance);
  return is_store;
}

/* Replace scalar calls from SLP node NODE with setting of their lhs to zero.
   For loop vectorization this is done in vectorizable_call, but for SLP
   it needs to be deferred until end of vect_schedule_slp, because multiple
   SLP instances may refer to the same scalar stmt.  */

static void
vect_remove_slp_scalar_calls (slp_tree node)
{
  gimple stmt, new_stmt;
  gimple_stmt_iterator gsi;
  int i;
  slp_tree child;
  tree lhs;
  stmt_vec_info stmt_info;

  if (!node)
    return;

  FOR_EACH_VEC_ELT (SLP_TREE_CHILDREN (node), i, child)
    vect_remove_slp_scalar_calls (child);

  FOR_EACH_VEC_ELT (SLP_TREE_SCALAR_STMTS (node), i, stmt)
    {
      if (!is_gimple_call (stmt) || gimple_bb (stmt) == NULL)
	continue;
      stmt_info = vinfo_for_stmt (stmt);
      if (stmt_info == NULL
	  || is_pattern_stmt_p (stmt_info)
	  || !PURE_SLP_STMT (stmt_info))
	continue;
      lhs = gimple_call_lhs (stmt);
      new_stmt = gimple_build_assign (lhs, build_zero_cst (TREE_TYPE (lhs)));
      set_vinfo_for_stmt (new_stmt, stmt_info);
      set_vinfo_for_stmt (stmt, NULL);
      STMT_VINFO_STMT (stmt_info) = new_stmt;
      gsi = gsi_for_stmt (stmt);
      gsi_replace (&gsi, new_stmt, false);
      SSA_NAME_DEF_STMT (gimple_assign_lhs (new_stmt)) = new_stmt;
    }
}

/* Generate vector code for all SLP instances in the loop/basic block.  */

bool
vect_schedule_slp (loop_vec_info loop_vinfo, bb_vec_info bb_vinfo)
{
  vec<slp_instance> slp_instances;
  slp_instance instance;
  unsigned int i, vf;
  bool is_store = false;

  if (loop_vinfo)
    {
      slp_instances = LOOP_VINFO_SLP_INSTANCES (loop_vinfo);
      vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
    }
  else
    {
      slp_instances = BB_VINFO_SLP_INSTANCES (bb_vinfo);
      vf = 1;
    }

  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    {
      /* Schedule the tree of INSTANCE.  */
      is_store = vect_schedule_slp_instance (SLP_INSTANCE_TREE (instance),
                                             instance, vf);
      if (dump_enabled_p ())
	dump_printf_loc (MSG_NOTE, vect_location,
                         "vectorizing stmts using SLP.\n");
    }

  FOR_EACH_VEC_ELT (slp_instances, i, instance)
    {
      slp_tree root = SLP_INSTANCE_TREE (instance);
      gimple store;
      unsigned int j;
      gimple_stmt_iterator gsi;

      /* Remove scalar call stmts.  Do not do this for basic-block
	 vectorization as not all uses may be vectorized.
	 ???  Why should this be necessary?  DCE should be able to
	 remove the stmts itself.
	 ???  For BB vectorization we can as well remove scalar
	 stmts starting from the SLP tree root if they have no
	 uses.  */
      if (loop_vinfo)
	vect_remove_slp_scalar_calls (root);

      for (j = 0; SLP_TREE_SCALAR_STMTS (root).iterate (j, &store)
                  && j < SLP_INSTANCE_GROUP_SIZE (instance); j++)
        {
          if (!STMT_VINFO_DATA_REF (vinfo_for_stmt (store)))
            break;

         if (is_pattern_stmt_p (vinfo_for_stmt (store)))
           store = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (store));
          /* Free the attached stmt_vec_info and remove the stmt.  */
          gsi = gsi_for_stmt (store);
	  unlink_stmt_vdef (store);
          gsi_remove (&gsi, true);
	  release_defs (store);
          free_stmt_vec_info (store);
        }
    }

  return is_store;
}


/* Vectorize the basic block.  */

void
vect_slp_transform_bb (basic_block bb)
{
  bb_vec_info bb_vinfo = vec_info_for_bb (bb);
  gimple_stmt_iterator si;

  gcc_assert (bb_vinfo);

  if (dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location, "SLPing BB\n");

  for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
    {
      gimple stmt = gsi_stmt (si);
      stmt_vec_info stmt_info;

      if (dump_enabled_p ())
        {
          dump_printf_loc (MSG_NOTE, vect_location,
                           "------>SLPing statement: ");
          dump_gimple_stmt (MSG_NOTE, TDF_SLIM, stmt, 0);
          dump_printf (MSG_NOTE, "\n");
        }

      stmt_info = vinfo_for_stmt (stmt);
      gcc_assert (stmt_info);

      /* Schedule all the SLP instances when the first SLP stmt is reached.  */
      if (STMT_SLP_TYPE (stmt_info))
        {
          vect_schedule_slp (NULL, bb_vinfo);
          break;
        }
    }

  if (dump_enabled_p ())
    dump_printf_loc (MSG_NOTE, vect_location,
		     "BASIC BLOCK VECTORIZED\n");

  destroy_bb_vec_info (bb_vinfo);
}