summaryrefslogtreecommitdiffstats
path: root/arch/riscv/net/bpf_jit_comp32.c
blob: 81de865f4c7c3592e6d730b2d3f723e9cacf09e7 (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
// SPDX-License-Identifier: GPL-2.0
/*
 * BPF JIT compiler for RV32G
 *
 * Copyright (c) 2020 Luke Nelson <luke.r.nels@gmail.com>
 * Copyright (c) 2020 Xi Wang <xi.wang@gmail.com>
 *
 * The code is based on the BPF JIT compiler for RV64G by Björn Töpel and
 * the BPF JIT compiler for 32-bit ARM by Shubham Bansal and Mircea Gherzan.
 */

#include <linux/bpf.h>
#include <linux/filter.h>
#include "bpf_jit.h"

/*
 * Stack layout during BPF program execution:
 *
 *                     high
 *     RV32 fp =>  +----------+
 *                 | saved ra |
 *                 | saved fp | RV32 callee-saved registers
 *                 |   ...    |
 *                 +----------+ <= (fp - 4 * NR_SAVED_REGISTERS)
 *                 |  hi(R6)  |
 *                 |  lo(R6)  |
 *                 |  hi(R7)  | JIT scratch space for BPF registers
 *                 |  lo(R7)  |
 *                 |   ...    |
 *  BPF_REG_FP =>  +----------+ <= (fp - 4 * NR_SAVED_REGISTERS
 *                 |          |        - 4 * BPF_JIT_SCRATCH_REGS)
 *                 |          |
 *                 |   ...    | BPF program stack
 *                 |          |
 *     RV32 sp =>  +----------+
 *                 |          |
 *                 |   ...    | Function call stack
 *                 |          |
 *                 +----------+
 *                     low
 */

enum {
	/* Stack layout - these are offsets from top of JIT scratch space. */
	BPF_R6_HI,
	BPF_R6_LO,
	BPF_R7_HI,
	BPF_R7_LO,
	BPF_R8_HI,
	BPF_R8_LO,
	BPF_R9_HI,
	BPF_R9_LO,
	BPF_AX_HI,
	BPF_AX_LO,
	/* Stack space for BPF_REG_6 through BPF_REG_9 and BPF_REG_AX. */
	BPF_JIT_SCRATCH_REGS,
};

/* Number of callee-saved registers stored to stack: ra, fp, s1--s7. */
#define NR_SAVED_REGISTERS	9

/* Offset from fp for BPF registers stored on stack. */
#define STACK_OFFSET(k)	(-4 - (4 * NR_SAVED_REGISTERS) - (4 * (k)))

#define TMP_REG_1	(MAX_BPF_JIT_REG + 0)
#define TMP_REG_2	(MAX_BPF_JIT_REG + 1)

#define RV_REG_TCC		RV_REG_T6
#define RV_REG_TCC_SAVED	RV_REG_S7

static const s8 bpf2rv32[][2] = {
	/* Return value from in-kernel function, and exit value from eBPF. */
	[BPF_REG_0] = {RV_REG_S2, RV_REG_S1},
	/* Arguments from eBPF program to in-kernel function. */
	[BPF_REG_1] = {RV_REG_A1, RV_REG_A0},
	[BPF_REG_2] = {RV_REG_A3, RV_REG_A2},
	[BPF_REG_3] = {RV_REG_A5, RV_REG_A4},
	[BPF_REG_4] = {RV_REG_A7, RV_REG_A6},
	[BPF_REG_5] = {RV_REG_S4, RV_REG_S3},
	/*
	 * Callee-saved registers that in-kernel function will preserve.
	 * Stored on the stack.
	 */
	[BPF_REG_6] = {STACK_OFFSET(BPF_R6_HI), STACK_OFFSET(BPF_R6_LO)},
	[BPF_REG_7] = {STACK_OFFSET(BPF_R7_HI), STACK_OFFSET(BPF_R7_LO)},
	[BPF_REG_8] = {STACK_OFFSET(BPF_R8_HI), STACK_OFFSET(BPF_R8_LO)},
	[BPF_REG_9] = {STACK_OFFSET(BPF_R9_HI), STACK_OFFSET(BPF_R9_LO)},
	/* Read-only frame pointer to access BPF stack. */
	[BPF_REG_FP] = {RV_REG_S6, RV_REG_S5},
	/* Temporary register for blinding constants. Stored on the stack. */
	[BPF_REG_AX] = {STACK_OFFSET(BPF_AX_HI), STACK_OFFSET(BPF_AX_LO)},
	/*
	 * Temporary registers used by the JIT to operate on registers stored
	 * on the stack. Save t0 and t1 to be used as temporaries in generated
	 * code.
	 */
	[TMP_REG_1] = {RV_REG_T3, RV_REG_T2},
	[TMP_REG_2] = {RV_REG_T5, RV_REG_T4},
};

static s8 hi(const s8 *r)
{
	return r[0];
}

static s8 lo(const s8 *r)
{
	return r[1];
}

static void emit_imm(const s8 rd, s32 imm, struct rv_jit_context *ctx)
{
	u32 upper = (imm + (1 << 11)) >> 12;
	u32 lower = imm & 0xfff;

	if (upper) {
		emit(rv_lui(rd, upper), ctx);
		emit(rv_addi(rd, rd, lower), ctx);
	} else {
		emit(rv_addi(rd, RV_REG_ZERO, lower), ctx);
	}
}

static void emit_imm32(const s8 *rd, s32 imm, struct rv_jit_context *ctx)
{
	/* Emit immediate into lower bits. */
	emit_imm(lo(rd), imm, ctx);

	/* Sign-extend into upper bits. */
	if (imm >= 0)
		emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
	else
		emit(rv_addi(hi(rd), RV_REG_ZERO, -1), ctx);
}

static void emit_imm64(const s8 *rd, s32 imm_hi, s32 imm_lo,
		       struct rv_jit_context *ctx)
{
	emit_imm(lo(rd), imm_lo, ctx);
	emit_imm(hi(rd), imm_hi, ctx);
}

static void __build_epilogue(bool is_tail_call, struct rv_jit_context *ctx)
{
	int stack_adjust = ctx->stack_size;
	const s8 *r0 = bpf2rv32[BPF_REG_0];

	/* Set return value if not tail call. */
	if (!is_tail_call) {
		emit(rv_addi(RV_REG_A0, lo(r0), 0), ctx);
		emit(rv_addi(RV_REG_A1, hi(r0), 0), ctx);
	}

	/* Restore callee-saved registers. */
	emit(rv_lw(RV_REG_RA, stack_adjust - 4, RV_REG_SP), ctx);
	emit(rv_lw(RV_REG_FP, stack_adjust - 8, RV_REG_SP), ctx);
	emit(rv_lw(RV_REG_S1, stack_adjust - 12, RV_REG_SP), ctx);
	emit(rv_lw(RV_REG_S2, stack_adjust - 16, RV_REG_SP), ctx);
	emit(rv_lw(RV_REG_S3, stack_adjust - 20, RV_REG_SP), ctx);
	emit(rv_lw(RV_REG_S4, stack_adjust - 24, RV_REG_SP), ctx);
	emit(rv_lw(RV_REG_S5, stack_adjust - 28, RV_REG_SP), ctx);
	emit(rv_lw(RV_REG_S6, stack_adjust - 32, RV_REG_SP), ctx);
	emit(rv_lw(RV_REG_S7, stack_adjust - 36, RV_REG_SP), ctx);

	emit(rv_addi(RV_REG_SP, RV_REG_SP, stack_adjust), ctx);

	if (is_tail_call) {
		/*
		 * goto *(t0 + 4);
		 * Skips first instruction of prologue which initializes tail
		 * call counter. Assumes t0 contains address of target program,
		 * see emit_bpf_tail_call.
		 */
		emit(rv_jalr(RV_REG_ZERO, RV_REG_T0, 4), ctx);
	} else {
		emit(rv_jalr(RV_REG_ZERO, RV_REG_RA, 0), ctx);
	}
}

static bool is_stacked(s8 reg)
{
	return reg < 0;
}

static const s8 *bpf_get_reg64(const s8 *reg, const s8 *tmp,
			       struct rv_jit_context *ctx)
{
	if (is_stacked(hi(reg))) {
		emit(rv_lw(hi(tmp), hi(reg), RV_REG_FP), ctx);
		emit(rv_lw(lo(tmp), lo(reg), RV_REG_FP), ctx);
		reg = tmp;
	}
	return reg;
}

static void bpf_put_reg64(const s8 *reg, const s8 *src,
			  struct rv_jit_context *ctx)
{
	if (is_stacked(hi(reg))) {
		emit(rv_sw(RV_REG_FP, hi(reg), hi(src)), ctx);
		emit(rv_sw(RV_REG_FP, lo(reg), lo(src)), ctx);
	}
}

static const s8 *bpf_get_reg32(const s8 *reg, const s8 *tmp,
			       struct rv_jit_context *ctx)
{
	if (is_stacked(lo(reg))) {
		emit(rv_lw(lo(tmp), lo(reg), RV_REG_FP), ctx);
		reg = tmp;
	}
	return reg;
}

static void bpf_put_reg32(const s8 *reg, const s8 *src,
			  struct rv_jit_context *ctx)
{
	if (is_stacked(lo(reg))) {
		emit(rv_sw(RV_REG_FP, lo(reg), lo(src)), ctx);
		if (!ctx->prog->aux->verifier_zext)
			emit(rv_sw(RV_REG_FP, hi(reg), RV_REG_ZERO), ctx);
	} else if (!ctx->prog->aux->verifier_zext) {
		emit(rv_addi(hi(reg), RV_REG_ZERO, 0), ctx);
	}
}

static void emit_jump_and_link(u8 rd, s32 rvoff, bool force_jalr,
			       struct rv_jit_context *ctx)
{
	s32 upper, lower;

	if (rvoff && is_21b_int(rvoff) && !force_jalr) {
		emit(rv_jal(rd, rvoff >> 1), ctx);
		return;
	}

	upper = (rvoff + (1 << 11)) >> 12;
	lower = rvoff & 0xfff;
	emit(rv_auipc(RV_REG_T1, upper), ctx);
	emit(rv_jalr(rd, RV_REG_T1, lower), ctx);
}

static void emit_alu_i64(const s8 *dst, s32 imm,
			 struct rv_jit_context *ctx, const u8 op)
{
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *rd = bpf_get_reg64(dst, tmp1, ctx);

	switch (op) {
	case BPF_MOV:
		emit_imm32(rd, imm, ctx);
		break;
	case BPF_AND:
		if (is_12b_int(imm)) {
			emit(rv_andi(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_and(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		if (imm >= 0)
			emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
		break;
	case BPF_OR:
		if (is_12b_int(imm)) {
			emit(rv_ori(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_or(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		if (imm < 0)
			emit(rv_ori(hi(rd), RV_REG_ZERO, -1), ctx);
		break;
	case BPF_XOR:
		if (is_12b_int(imm)) {
			emit(rv_xori(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_xor(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		if (imm < 0)
			emit(rv_xori(hi(rd), hi(rd), -1), ctx);
		break;
	case BPF_LSH:
		if (imm >= 32) {
			emit(rv_slli(hi(rd), lo(rd), imm - 32), ctx);
			emit(rv_addi(lo(rd), RV_REG_ZERO, 0), ctx);
		} else if (imm == 0) {
			/* Do nothing. */
		} else {
			emit(rv_srli(RV_REG_T0, lo(rd), 32 - imm), ctx);
			emit(rv_slli(hi(rd), hi(rd), imm), ctx);
			emit(rv_or(hi(rd), RV_REG_T0, hi(rd)), ctx);
			emit(rv_slli(lo(rd), lo(rd), imm), ctx);
		}
		break;
	case BPF_RSH:
		if (imm >= 32) {
			emit(rv_srli(lo(rd), hi(rd), imm - 32), ctx);
			emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
		} else if (imm == 0) {
			/* Do nothing. */
		} else {
			emit(rv_slli(RV_REG_T0, hi(rd), 32 - imm), ctx);
			emit(rv_srli(lo(rd), lo(rd), imm), ctx);
			emit(rv_or(lo(rd), RV_REG_T0, lo(rd)), ctx);
			emit(rv_srli(hi(rd), hi(rd), imm), ctx);
		}
		break;
	case BPF_ARSH:
		if (imm >= 32) {
			emit(rv_srai(lo(rd), hi(rd), imm - 32), ctx);
			emit(rv_srai(hi(rd), hi(rd), 31), ctx);
		} else if (imm == 0) {
			/* Do nothing. */
		} else {
			emit(rv_slli(RV_REG_T0, hi(rd), 32 - imm), ctx);
			emit(rv_srli(lo(rd), lo(rd), imm), ctx);
			emit(rv_or(lo(rd), RV_REG_T0, lo(rd)), ctx);
			emit(rv_srai(hi(rd), hi(rd), imm), ctx);
		}
		break;
	}

	bpf_put_reg64(dst, rd, ctx);
}

static void emit_alu_i32(const s8 *dst, s32 imm,
			 struct rv_jit_context *ctx, const u8 op)
{
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *rd = bpf_get_reg32(dst, tmp1, ctx);

	switch (op) {
	case BPF_MOV:
		emit_imm(lo(rd), imm, ctx);
		break;
	case BPF_ADD:
		if (is_12b_int(imm)) {
			emit(rv_addi(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_add(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		break;
	case BPF_SUB:
		if (is_12b_int(-imm)) {
			emit(rv_addi(lo(rd), lo(rd), -imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_sub(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		break;
	case BPF_AND:
		if (is_12b_int(imm)) {
			emit(rv_andi(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_and(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		break;
	case BPF_OR:
		if (is_12b_int(imm)) {
			emit(rv_ori(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_or(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		break;
	case BPF_XOR:
		if (is_12b_int(imm)) {
			emit(rv_xori(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_xor(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		break;
	case BPF_LSH:
		if (is_12b_int(imm)) {
			emit(rv_slli(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_sll(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		break;
	case BPF_RSH:
		if (is_12b_int(imm)) {
			emit(rv_srli(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_srl(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		break;
	case BPF_ARSH:
		if (is_12b_int(imm)) {
			emit(rv_srai(lo(rd), lo(rd), imm), ctx);
		} else {
			emit_imm(RV_REG_T0, imm, ctx);
			emit(rv_sra(lo(rd), lo(rd), RV_REG_T0), ctx);
		}
		break;
	}

	bpf_put_reg32(dst, rd, ctx);
}

static void emit_alu_r64(const s8 *dst, const s8 *src,
			 struct rv_jit_context *ctx, const u8 op)
{
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *tmp2 = bpf2rv32[TMP_REG_2];
	const s8 *rd = bpf_get_reg64(dst, tmp1, ctx);
	const s8 *rs = bpf_get_reg64(src, tmp2, ctx);

	switch (op) {
	case BPF_MOV:
		emit(rv_addi(lo(rd), lo(rs), 0), ctx);
		emit(rv_addi(hi(rd), hi(rs), 0), ctx);
		break;
	case BPF_ADD:
		if (rd == rs) {
			emit(rv_srli(RV_REG_T0, lo(rd), 31), ctx);
			emit(rv_slli(hi(rd), hi(rd), 1), ctx);
			emit(rv_or(hi(rd), RV_REG_T0, hi(rd)), ctx);
			emit(rv_slli(lo(rd), lo(rd), 1), ctx);
		} else {
			emit(rv_add(lo(rd), lo(rd), lo(rs)), ctx);
			emit(rv_sltu(RV_REG_T0, lo(rd), lo(rs)), ctx);
			emit(rv_add(hi(rd), hi(rd), hi(rs)), ctx);
			emit(rv_add(hi(rd), hi(rd), RV_REG_T0), ctx);
		}
		break;
	case BPF_SUB:
		emit(rv_sub(RV_REG_T1, hi(rd), hi(rs)), ctx);
		emit(rv_sltu(RV_REG_T0, lo(rd), lo(rs)), ctx);
		emit(rv_sub(hi(rd), RV_REG_T1, RV_REG_T0), ctx);
		emit(rv_sub(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_AND:
		emit(rv_and(lo(rd), lo(rd), lo(rs)), ctx);
		emit(rv_and(hi(rd), hi(rd), hi(rs)), ctx);
		break;
	case BPF_OR:
		emit(rv_or(lo(rd), lo(rd), lo(rs)), ctx);
		emit(rv_or(hi(rd), hi(rd), hi(rs)), ctx);
		break;
	case BPF_XOR:
		emit(rv_xor(lo(rd), lo(rd), lo(rs)), ctx);
		emit(rv_xor(hi(rd), hi(rd), hi(rs)), ctx);
		break;
	case BPF_MUL:
		emit(rv_mul(RV_REG_T0, hi(rs), lo(rd)), ctx);
		emit(rv_mul(hi(rd), hi(rd), lo(rs)), ctx);
		emit(rv_mulhu(RV_REG_T1, lo(rd), lo(rs)), ctx);
		emit(rv_add(hi(rd), hi(rd), RV_REG_T0), ctx);
		emit(rv_mul(lo(rd), lo(rd), lo(rs)), ctx);
		emit(rv_add(hi(rd), hi(rd), RV_REG_T1), ctx);
		break;
	case BPF_LSH:
		emit(rv_addi(RV_REG_T0, lo(rs), -32), ctx);
		emit(rv_blt(RV_REG_T0, RV_REG_ZERO, 8), ctx);
		emit(rv_sll(hi(rd), lo(rd), RV_REG_T0), ctx);
		emit(rv_addi(lo(rd), RV_REG_ZERO, 0), ctx);
		emit(rv_jal(RV_REG_ZERO, 16), ctx);
		emit(rv_addi(RV_REG_T1, RV_REG_ZERO, 31), ctx);
		emit(rv_srli(RV_REG_T0, lo(rd), 1), ctx);
		emit(rv_sub(RV_REG_T1, RV_REG_T1, lo(rs)), ctx);
		emit(rv_srl(RV_REG_T0, RV_REG_T0, RV_REG_T1), ctx);
		emit(rv_sll(hi(rd), hi(rd), lo(rs)), ctx);
		emit(rv_or(hi(rd), RV_REG_T0, hi(rd)), ctx);
		emit(rv_sll(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_RSH:
		emit(rv_addi(RV_REG_T0, lo(rs), -32), ctx);
		emit(rv_blt(RV_REG_T0, RV_REG_ZERO, 8), ctx);
		emit(rv_srl(lo(rd), hi(rd), RV_REG_T0), ctx);
		emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
		emit(rv_jal(RV_REG_ZERO, 16), ctx);
		emit(rv_addi(RV_REG_T1, RV_REG_ZERO, 31), ctx);
		emit(rv_slli(RV_REG_T0, hi(rd), 1), ctx);
		emit(rv_sub(RV_REG_T1, RV_REG_T1, lo(rs)), ctx);
		emit(rv_sll(RV_REG_T0, RV_REG_T0, RV_REG_T1), ctx);
		emit(rv_srl(lo(rd), lo(rd), lo(rs)), ctx);
		emit(rv_or(lo(rd), RV_REG_T0, lo(rd)), ctx);
		emit(rv_srl(hi(rd), hi(rd), lo(rs)), ctx);
		break;
	case BPF_ARSH:
		emit(rv_addi(RV_REG_T0, lo(rs), -32), ctx);
		emit(rv_blt(RV_REG_T0, RV_REG_ZERO, 8), ctx);
		emit(rv_sra(lo(rd), hi(rd), RV_REG_T0), ctx);
		emit(rv_srai(hi(rd), hi(rd), 31), ctx);
		emit(rv_jal(RV_REG_ZERO, 16), ctx);
		emit(rv_addi(RV_REG_T1, RV_REG_ZERO, 31), ctx);
		emit(rv_slli(RV_REG_T0, hi(rd), 1), ctx);
		emit(rv_sub(RV_REG_T1, RV_REG_T1, lo(rs)), ctx);
		emit(rv_sll(RV_REG_T0, RV_REG_T0, RV_REG_T1), ctx);
		emit(rv_srl(lo(rd), lo(rd), lo(rs)), ctx);
		emit(rv_or(lo(rd), RV_REG_T0, lo(rd)), ctx);
		emit(rv_sra(hi(rd), hi(rd), lo(rs)), ctx);
		break;
	case BPF_NEG:
		emit(rv_sub(lo(rd), RV_REG_ZERO, lo(rd)), ctx);
		emit(rv_sltu(RV_REG_T0, RV_REG_ZERO, lo(rd)), ctx);
		emit(rv_sub(hi(rd), RV_REG_ZERO, hi(rd)), ctx);
		emit(rv_sub(hi(rd), hi(rd), RV_REG_T0), ctx);
		break;
	}

	bpf_put_reg64(dst, rd, ctx);
}

static void emit_alu_r32(const s8 *dst, const s8 *src,
			 struct rv_jit_context *ctx, const u8 op)
{
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *tmp2 = bpf2rv32[TMP_REG_2];
	const s8 *rd = bpf_get_reg32(dst, tmp1, ctx);
	const s8 *rs = bpf_get_reg32(src, tmp2, ctx);

	switch (op) {
	case BPF_MOV:
		emit(rv_addi(lo(rd), lo(rs), 0), ctx);
		break;
	case BPF_ADD:
		emit(rv_add(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_SUB:
		emit(rv_sub(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_AND:
		emit(rv_and(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_OR:
		emit(rv_or(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_XOR:
		emit(rv_xor(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_MUL:
		emit(rv_mul(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_DIV:
		emit(rv_divu(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_MOD:
		emit(rv_remu(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_LSH:
		emit(rv_sll(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_RSH:
		emit(rv_srl(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_ARSH:
		emit(rv_sra(lo(rd), lo(rd), lo(rs)), ctx);
		break;
	case BPF_NEG:
		emit(rv_sub(lo(rd), RV_REG_ZERO, lo(rd)), ctx);
		break;
	}

	bpf_put_reg32(dst, rd, ctx);
}

static int emit_branch_r64(const s8 *src1, const s8 *src2, s32 rvoff,
			   struct rv_jit_context *ctx, const u8 op)
{
	int e, s = ctx->ninsns;
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *tmp2 = bpf2rv32[TMP_REG_2];

	const s8 *rs1 = bpf_get_reg64(src1, tmp1, ctx);
	const s8 *rs2 = bpf_get_reg64(src2, tmp2, ctx);

	/*
	 * NO_JUMP skips over the rest of the instructions and the
	 * emit_jump_and_link, meaning the BPF branch is not taken.
	 * JUMP skips directly to the emit_jump_and_link, meaning
	 * the BPF branch is taken.
	 *
	 * The fallthrough case results in the BPF branch being taken.
	 */
#define NO_JUMP(idx) (6 + (2 * (idx)))
#define JUMP(idx) (2 + (2 * (idx)))

	switch (op) {
	case BPF_JEQ:
		emit(rv_bne(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bne(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JGT:
		emit(rv_bgtu(hi(rs1), hi(rs2), JUMP(2)), ctx);
		emit(rv_bltu(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bleu(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JLT:
		emit(rv_bltu(hi(rs1), hi(rs2), JUMP(2)), ctx);
		emit(rv_bgtu(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bgeu(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JGE:
		emit(rv_bgtu(hi(rs1), hi(rs2), JUMP(2)), ctx);
		emit(rv_bltu(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bltu(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JLE:
		emit(rv_bltu(hi(rs1), hi(rs2), JUMP(2)), ctx);
		emit(rv_bgtu(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bgtu(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JNE:
		emit(rv_bne(hi(rs1), hi(rs2), JUMP(1)), ctx);
		emit(rv_beq(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JSGT:
		emit(rv_bgt(hi(rs1), hi(rs2), JUMP(2)), ctx);
		emit(rv_blt(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bleu(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JSLT:
		emit(rv_blt(hi(rs1), hi(rs2), JUMP(2)), ctx);
		emit(rv_bgt(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bgeu(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JSGE:
		emit(rv_bgt(hi(rs1), hi(rs2), JUMP(2)), ctx);
		emit(rv_blt(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bltu(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JSLE:
		emit(rv_blt(hi(rs1), hi(rs2), JUMP(2)), ctx);
		emit(rv_bgt(hi(rs1), hi(rs2), NO_JUMP(1)), ctx);
		emit(rv_bgtu(lo(rs1), lo(rs2), NO_JUMP(0)), ctx);
		break;
	case BPF_JSET:
		emit(rv_and(RV_REG_T0, hi(rs1), hi(rs2)), ctx);
		emit(rv_bne(RV_REG_T0, RV_REG_ZERO, JUMP(2)), ctx);
		emit(rv_and(RV_REG_T0, lo(rs1), lo(rs2)), ctx);
		emit(rv_beq(RV_REG_T0, RV_REG_ZERO, NO_JUMP(0)), ctx);
		break;
	}

#undef NO_JUMP
#undef JUMP

	e = ctx->ninsns;
	/* Adjust for extra insns. */
	rvoff -= ninsns_rvoff(e - s);
	emit_jump_and_link(RV_REG_ZERO, rvoff, true, ctx);
	return 0;
}

static int emit_bcc(u8 op, u8 rd, u8 rs, int rvoff, struct rv_jit_context *ctx)
{
	int e, s = ctx->ninsns;
	bool far = false;
	int off;

	if (op == BPF_JSET) {
		/*
		 * BPF_JSET is a special case: it has no inverse so we always
		 * treat it as a far branch.
		 */
		far = true;
	} else if (!is_13b_int(rvoff)) {
		op = invert_bpf_cond(op);
		far = true;
	}

	/*
	 * For a far branch, the condition is negated and we jump over the
	 * branch itself, and the two instructions from emit_jump_and_link.
	 * For a near branch, just use rvoff.
	 */
	off = far ? 6 : (rvoff >> 1);

	switch (op) {
	case BPF_JEQ:
		emit(rv_beq(rd, rs, off), ctx);
		break;
	case BPF_JGT:
		emit(rv_bgtu(rd, rs, off), ctx);
		break;
	case BPF_JLT:
		emit(rv_bltu(rd, rs, off), ctx);
		break;
	case BPF_JGE:
		emit(rv_bgeu(rd, rs, off), ctx);
		break;
	case BPF_JLE:
		emit(rv_bleu(rd, rs, off), ctx);
		break;
	case BPF_JNE:
		emit(rv_bne(rd, rs, off), ctx);
		break;
	case BPF_JSGT:
		emit(rv_bgt(rd, rs, off), ctx);
		break;
	case BPF_JSLT:
		emit(rv_blt(rd, rs, off), ctx);
		break;
	case BPF_JSGE:
		emit(rv_bge(rd, rs, off), ctx);
		break;
	case BPF_JSLE:
		emit(rv_ble(rd, rs, off), ctx);
		break;
	case BPF_JSET:
		emit(rv_and(RV_REG_T0, rd, rs), ctx);
		emit(rv_beq(RV_REG_T0, RV_REG_ZERO, off), ctx);
		break;
	}

	if (far) {
		e = ctx->ninsns;
		/* Adjust for extra insns. */
		rvoff -= ninsns_rvoff(e - s);
		emit_jump_and_link(RV_REG_ZERO, rvoff, true, ctx);
	}
	return 0;
}

static int emit_branch_r32(const s8 *src1, const s8 *src2, s32 rvoff,
			   struct rv_jit_context *ctx, const u8 op)
{
	int e, s = ctx->ninsns;
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *tmp2 = bpf2rv32[TMP_REG_2];

	const s8 *rs1 = bpf_get_reg32(src1, tmp1, ctx);
	const s8 *rs2 = bpf_get_reg32(src2, tmp2, ctx);

	e = ctx->ninsns;
	/* Adjust for extra insns. */
	rvoff -= ninsns_rvoff(e - s);

	if (emit_bcc(op, lo(rs1), lo(rs2), rvoff, ctx))
		return -1;

	return 0;
}

static void emit_call(bool fixed, u64 addr, struct rv_jit_context *ctx)
{
	const s8 *r0 = bpf2rv32[BPF_REG_0];
	const s8 *r5 = bpf2rv32[BPF_REG_5];
	u32 upper = ((u32)addr + (1 << 11)) >> 12;
	u32 lower = addr & 0xfff;

	/* R1-R4 already in correct registers---need to push R5 to stack. */
	emit(rv_addi(RV_REG_SP, RV_REG_SP, -16), ctx);
	emit(rv_sw(RV_REG_SP, 0, lo(r5)), ctx);
	emit(rv_sw(RV_REG_SP, 4, hi(r5)), ctx);

	/* Backup TCC. */
	emit(rv_addi(RV_REG_TCC_SAVED, RV_REG_TCC, 0), ctx);

	/*
	 * Use lui/jalr pair to jump to absolute address. Don't use emit_imm as
	 * the number of emitted instructions should not depend on the value of
	 * addr.
	 */
	emit(rv_lui(RV_REG_T1, upper), ctx);
	emit(rv_jalr(RV_REG_RA, RV_REG_T1, lower), ctx);

	/* Restore TCC. */
	emit(rv_addi(RV_REG_TCC, RV_REG_TCC_SAVED, 0), ctx);

	/* Set return value and restore stack. */
	emit(rv_addi(lo(r0), RV_REG_A0, 0), ctx);
	emit(rv_addi(hi(r0), RV_REG_A1, 0), ctx);
	emit(rv_addi(RV_REG_SP, RV_REG_SP, 16), ctx);
}

static int emit_bpf_tail_call(int insn, struct rv_jit_context *ctx)
{
	/*
	 * R1 -> &ctx
	 * R2 -> &array
	 * R3 -> index
	 */
	int tc_ninsn, off, start_insn = ctx->ninsns;
	const s8 *arr_reg = bpf2rv32[BPF_REG_2];
	const s8 *idx_reg = bpf2rv32[BPF_REG_3];

	tc_ninsn = insn ? ctx->offset[insn] - ctx->offset[insn - 1] :
		ctx->offset[0];

	/* max_entries = array->map.max_entries; */
	off = offsetof(struct bpf_array, map.max_entries);
	if (is_12b_check(off, insn))
		return -1;
	emit(rv_lw(RV_REG_T1, off, lo(arr_reg)), ctx);

	/*
	 * if (index >= max_entries)
	 *   goto out;
	 */
	off = ninsns_rvoff(tc_ninsn - (ctx->ninsns - start_insn));
	emit_bcc(BPF_JGE, lo(idx_reg), RV_REG_T1, off, ctx);

	/*
	 * temp_tcc = tcc - 1;
	 * if (tcc < 0)
	 *   goto out;
	 */
	emit(rv_addi(RV_REG_T1, RV_REG_TCC, -1), ctx);
	off = ninsns_rvoff(tc_ninsn - (ctx->ninsns - start_insn));
	emit_bcc(BPF_JSLT, RV_REG_TCC, RV_REG_ZERO, off, ctx);

	/*
	 * prog = array->ptrs[index];
	 * if (!prog)
	 *   goto out;
	 */
	emit(rv_slli(RV_REG_T0, lo(idx_reg), 2), ctx);
	emit(rv_add(RV_REG_T0, RV_REG_T0, lo(arr_reg)), ctx);
	off = offsetof(struct bpf_array, ptrs);
	if (is_12b_check(off, insn))
		return -1;
	emit(rv_lw(RV_REG_T0, off, RV_REG_T0), ctx);
	off = ninsns_rvoff(tc_ninsn - (ctx->ninsns - start_insn));
	emit_bcc(BPF_JEQ, RV_REG_T0, RV_REG_ZERO, off, ctx);

	/*
	 * tcc = temp_tcc;
	 * goto *(prog->bpf_func + 4);
	 */
	off = offsetof(struct bpf_prog, bpf_func);
	if (is_12b_check(off, insn))
		return -1;
	emit(rv_lw(RV_REG_T0, off, RV_REG_T0), ctx);
	emit(rv_addi(RV_REG_TCC, RV_REG_T1, 0), ctx);
	/* Epilogue jumps to *(t0 + 4). */
	__build_epilogue(true, ctx);
	return 0;
}

static int emit_load_r64(const s8 *dst, const s8 *src, s16 off,
			 struct rv_jit_context *ctx, const u8 size)
{
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *tmp2 = bpf2rv32[TMP_REG_2];
	const s8 *rd = bpf_get_reg64(dst, tmp1, ctx);
	const s8 *rs = bpf_get_reg64(src, tmp2, ctx);

	emit_imm(RV_REG_T0, off, ctx);
	emit(rv_add(RV_REG_T0, RV_REG_T0, lo(rs)), ctx);

	switch (size) {
	case BPF_B:
		emit(rv_lbu(lo(rd), 0, RV_REG_T0), ctx);
		if (!ctx->prog->aux->verifier_zext)
			emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
		break;
	case BPF_H:
		emit(rv_lhu(lo(rd), 0, RV_REG_T0), ctx);
		if (!ctx->prog->aux->verifier_zext)
			emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
		break;
	case BPF_W:
		emit(rv_lw(lo(rd), 0, RV_REG_T0), ctx);
		if (!ctx->prog->aux->verifier_zext)
			emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
		break;
	case BPF_DW:
		emit(rv_lw(lo(rd), 0, RV_REG_T0), ctx);
		emit(rv_lw(hi(rd), 4, RV_REG_T0), ctx);
		break;
	}

	bpf_put_reg64(dst, rd, ctx);
	return 0;
}

static int emit_store_r64(const s8 *dst, const s8 *src, s16 off,
			  struct rv_jit_context *ctx, const u8 size,
			  const u8 mode)
{
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *tmp2 = bpf2rv32[TMP_REG_2];
	const s8 *rd = bpf_get_reg64(dst, tmp1, ctx);
	const s8 *rs = bpf_get_reg64(src, tmp2, ctx);

	if (mode == BPF_ATOMIC && size != BPF_W)
		return -1;

	emit_imm(RV_REG_T0, off, ctx);
	emit(rv_add(RV_REG_T0, RV_REG_T0, lo(rd)), ctx);

	switch (size) {
	case BPF_B:
		emit(rv_sb(RV_REG_T0, 0, lo(rs)), ctx);
		break;
	case BPF_H:
		emit(rv_sh(RV_REG_T0, 0, lo(rs)), ctx);
		break;
	case BPF_W:
		switch (mode) {
		case BPF_MEM:
			emit(rv_sw(RV_REG_T0, 0, lo(rs)), ctx);
			break;
		case BPF_ATOMIC: /* Only BPF_ADD supported */
			emit(rv_amoadd_w(RV_REG_ZERO, lo(rs), RV_REG_T0, 0, 0),
			     ctx);
			break;
		}
		break;
	case BPF_DW:
		emit(rv_sw(RV_REG_T0, 0, lo(rs)), ctx);
		emit(rv_sw(RV_REG_T0, 4, hi(rs)), ctx);
		break;
	}

	return 0;
}

static void emit_rev16(const s8 rd, struct rv_jit_context *ctx)
{
	emit(rv_slli(rd, rd, 16), ctx);
	emit(rv_slli(RV_REG_T1, rd, 8), ctx);
	emit(rv_srli(rd, rd, 8), ctx);
	emit(rv_add(RV_REG_T1, rd, RV_REG_T1), ctx);
	emit(rv_srli(rd, RV_REG_T1, 16), ctx);
}

static void emit_rev32(const s8 rd, struct rv_jit_context *ctx)
{
	emit(rv_addi(RV_REG_T1, RV_REG_ZERO, 0), ctx);
	emit(rv_andi(RV_REG_T0, rd, 255), ctx);
	emit(rv_add(RV_REG_T1, RV_REG_T1, RV_REG_T0), ctx);
	emit(rv_slli(RV_REG_T1, RV_REG_T1, 8), ctx);
	emit(rv_srli(rd, rd, 8), ctx);
	emit(rv_andi(RV_REG_T0, rd, 255), ctx);
	emit(rv_add(RV_REG_T1, RV_REG_T1, RV_REG_T0), ctx);
	emit(rv_slli(RV_REG_T1, RV_REG_T1, 8), ctx);
	emit(rv_srli(rd, rd, 8), ctx);
	emit(rv_andi(RV_REG_T0, rd, 255), ctx);
	emit(rv_add(RV_REG_T1, RV_REG_T1, RV_REG_T0), ctx);
	emit(rv_slli(RV_REG_T1, RV_REG_T1, 8), ctx);
	emit(rv_srli(rd, rd, 8), ctx);
	emit(rv_andi(RV_REG_T0, rd, 255), ctx);
	emit(rv_add(RV_REG_T1, RV_REG_T1, RV_REG_T0), ctx);
	emit(rv_addi(rd, RV_REG_T1, 0), ctx);
}

static void emit_zext64(const s8 *dst, struct rv_jit_context *ctx)
{
	const s8 *rd;
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];

	rd = bpf_get_reg64(dst, tmp1, ctx);
	emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
	bpf_put_reg64(dst, rd, ctx);
}

int bpf_jit_emit_insn(const struct bpf_insn *insn, struct rv_jit_context *ctx,
		      bool extra_pass)
{
	bool is64 = BPF_CLASS(insn->code) == BPF_ALU64 ||
		BPF_CLASS(insn->code) == BPF_JMP;
	int s, e, rvoff, i = insn - ctx->prog->insnsi;
	u8 code = insn->code;
	s16 off = insn->off;
	s32 imm = insn->imm;

	const s8 *dst = bpf2rv32[insn->dst_reg];
	const s8 *src = bpf2rv32[insn->src_reg];
	const s8 *tmp1 = bpf2rv32[TMP_REG_1];
	const s8 *tmp2 = bpf2rv32[TMP_REG_2];

	switch (code) {
	case BPF_ALU64 | BPF_MOV | BPF_X:

	case BPF_ALU64 | BPF_ADD | BPF_X:
	case BPF_ALU64 | BPF_ADD | BPF_K:

	case BPF_ALU64 | BPF_SUB | BPF_X:
	case BPF_ALU64 | BPF_SUB | BPF_K:

	case BPF_ALU64 | BPF_AND | BPF_X:
	case BPF_ALU64 | BPF_OR | BPF_X:
	case BPF_ALU64 | BPF_XOR | BPF_X:

	case BPF_ALU64 | BPF_MUL | BPF_X:
	case BPF_ALU64 | BPF_MUL | BPF_K:

	case BPF_ALU64 | BPF_LSH | BPF_X:
	case BPF_ALU64 | BPF_RSH | BPF_X:
	case BPF_ALU64 | BPF_ARSH | BPF_X:
		if (BPF_SRC(code) == BPF_K) {
			emit_imm32(tmp2, imm, ctx);
			src = tmp2;
		}
		emit_alu_r64(dst, src, ctx, BPF_OP(code));
		break;

	case BPF_ALU64 | BPF_NEG:
		emit_alu_r64(dst, tmp2, ctx, BPF_OP(code));
		break;

	case BPF_ALU64 | BPF_DIV | BPF_X:
	case BPF_ALU64 | BPF_DIV | BPF_K:
	case BPF_ALU64 | BPF_MOD | BPF_X:
	case BPF_ALU64 | BPF_MOD | BPF_K:
		goto notsupported;

	case BPF_ALU64 | BPF_MOV | BPF_K:
	case BPF_ALU64 | BPF_AND | BPF_K:
	case BPF_ALU64 | BPF_OR | BPF_K:
	case BPF_ALU64 | BPF_XOR | BPF_K:
	case BPF_ALU64 | BPF_LSH | BPF_K:
	case BPF_ALU64 | BPF_RSH | BPF_K:
	case BPF_ALU64 | BPF_ARSH | BPF_K:
		emit_alu_i64(dst, imm, ctx, BPF_OP(code));
		break;

	case BPF_ALU | BPF_MOV | BPF_X:
		if (imm == 1) {
			/* Special mov32 for zext. */
			emit_zext64(dst, ctx);
			break;
		}
		fallthrough;

	case BPF_ALU | BPF_ADD | BPF_X:
	case BPF_ALU | BPF_SUB | BPF_X:
	case BPF_ALU | BPF_AND | BPF_X:
	case BPF_ALU | BPF_OR | BPF_X:
	case BPF_ALU | BPF_XOR | BPF_X:

	case BPF_ALU | BPF_MUL | BPF_X:
	case BPF_ALU | BPF_MUL | BPF_K:

	case BPF_ALU | BPF_DIV | BPF_X:
	case BPF_ALU | BPF_DIV | BPF_K:

	case BPF_ALU | BPF_MOD | BPF_X:
	case BPF_ALU | BPF_MOD | BPF_K:

	case BPF_ALU | BPF_LSH | BPF_X:
	case BPF_ALU | BPF_RSH | BPF_X:
	case BPF_ALU | BPF_ARSH | BPF_X:
		if (BPF_SRC(code) == BPF_K) {
			emit_imm32(tmp2, imm, ctx);
			src = tmp2;
		}
		emit_alu_r32(dst, src, ctx, BPF_OP(code));
		break;

	case BPF_ALU | BPF_MOV | BPF_K:
	case BPF_ALU | BPF_ADD | BPF_K:
	case BPF_ALU | BPF_SUB | BPF_K:
	case BPF_ALU | BPF_AND | BPF_K:
	case BPF_ALU | BPF_OR | BPF_K:
	case BPF_ALU | BPF_XOR | BPF_K:
	case BPF_ALU | BPF_LSH | BPF_K:
	case BPF_ALU | BPF_RSH | BPF_K:
	case BPF_ALU | BPF_ARSH | BPF_K:
		/*
		 * mul,div,mod are handled in the BPF_X case since there are
		 * no RISC-V I-type equivalents.
		 */
		emit_alu_i32(dst, imm, ctx, BPF_OP(code));
		break;

	case BPF_ALU | BPF_NEG:
		/*
		 * src is ignored---choose tmp2 as a dummy register since it
		 * is not on the stack.
		 */
		emit_alu_r32(dst, tmp2, ctx, BPF_OP(code));
		break;

	case BPF_ALU | BPF_END | BPF_FROM_LE:
	{
		const s8 *rd = bpf_get_reg64(dst, tmp1, ctx);

		switch (imm) {
		case 16:
			emit(rv_slli(lo(rd), lo(rd), 16), ctx);
			emit(rv_srli(lo(rd), lo(rd), 16), ctx);
			fallthrough;
		case 32:
			if (!ctx->prog->aux->verifier_zext)
				emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
			break;
		case 64:
			/* Do nothing. */
			break;
		default:
			pr_err("bpf-jit: BPF_END imm %d invalid\n", imm);
			return -1;
		}

		bpf_put_reg64(dst, rd, ctx);
		break;
	}

	case BPF_ALU | BPF_END | BPF_FROM_BE:
	{
		const s8 *rd = bpf_get_reg64(dst, tmp1, ctx);

		switch (imm) {
		case 16:
			emit_rev16(lo(rd), ctx);
			if (!ctx->prog->aux->verifier_zext)
				emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
			break;
		case 32:
			emit_rev32(lo(rd), ctx);
			if (!ctx->prog->aux->verifier_zext)
				emit(rv_addi(hi(rd), RV_REG_ZERO, 0), ctx);
			break;
		case 64:
			/* Swap upper and lower halves. */
			emit(rv_addi(RV_REG_T0, lo(rd), 0), ctx);
			emit(rv_addi(lo(rd), hi(rd), 0), ctx);
			emit(rv_addi(hi(rd), RV_REG_T0, 0), ctx);

			/* Swap each half. */
			emit_rev32(lo(rd), ctx);
			emit_rev32(hi(rd), ctx);
			break;
		default:
			pr_err("bpf-jit: BPF_END imm %d invalid\n", imm);
			return -1;
		}

		bpf_put_reg64(dst, rd, ctx);
		break;
	}

	case BPF_JMP | BPF_JA:
		rvoff = rv_offset(i, off, ctx);
		emit_jump_and_link(RV_REG_ZERO, rvoff, false, ctx);
		break;

	case BPF_JMP | BPF_CALL:
	{
		bool fixed;
		int ret;
		u64 addr;

		ret = bpf_jit_get_func_addr(ctx->prog, insn, extra_pass, &addr,
					    &fixed);
		if (ret < 0)
			return ret;
		emit_call(fixed, addr, ctx);
		break;
	}

	case BPF_JMP | BPF_TAIL_CALL:
		if (emit_bpf_tail_call(i, ctx))
			return -1;
		break;

	case BPF_JMP | BPF_JEQ | BPF_X:
	case BPF_JMP | BPF_JEQ | BPF_K:
	case BPF_JMP32 | BPF_JEQ | BPF_X:
	case BPF_JMP32 | BPF_JEQ | BPF_K:

	case BPF_JMP | BPF_JNE | BPF_X:
	case BPF_JMP | BPF_JNE | BPF_K:
	case BPF_JMP32 | BPF_JNE | BPF_X:
	case BPF_JMP32 | BPF_JNE | BPF_K:

	case BPF_JMP | BPF_JLE | BPF_X:
	case BPF_JMP | BPF_JLE | BPF_K:
	case BPF_JMP32 | BPF_JLE | BPF_X:
	case BPF_JMP32 | BPF_JLE | BPF_K:

	case BPF_JMP | BPF_JLT | BPF_X:
	case BPF_JMP | BPF_JLT | BPF_K:
	case BPF_JMP32 | BPF_JLT | BPF_X:
	case BPF_JMP32 | BPF_JLT | BPF_K:

	case BPF_JMP | BPF_JGE | BPF_X:
	case BPF_JMP | BPF_JGE | BPF_K:
	case BPF_JMP32 | BPF_JGE | BPF_X:
	case BPF_JMP32 | BPF_JGE | BPF_K:

	case BPF_JMP | BPF_JGT | BPF_X:
	case BPF_JMP | BPF_JGT | BPF_K:
	case BPF_JMP32 | BPF_JGT | BPF_X:
	case BPF_JMP32 | BPF_JGT | BPF_K:

	case BPF_JMP | BPF_JSLE | BPF_X:
	case BPF_JMP | BPF_JSLE | BPF_K:
	case BPF_JMP32 | BPF_JSLE | BPF_X:
	case BPF_JMP32 | BPF_JSLE | BPF_K:

	case BPF_JMP | BPF_JSLT | BPF_X:
	case BPF_JMP | BPF_JSLT | BPF_K:
	case BPF_JMP32 | BPF_JSLT | BPF_X:
	case BPF_JMP32 | BPF_JSLT | BPF_K:

	case BPF_JMP | BPF_JSGE | BPF_X:
	case BPF_JMP | BPF_JSGE | BPF_K:
	case BPF_JMP32 | BPF_JSGE | BPF_X:
	case BPF_JMP32 | BPF_JSGE | BPF_K:

	case BPF_JMP | BPF_JSGT | BPF_X:
	case BPF_JMP | BPF_JSGT | BPF_K:
	case BPF_JMP32 | BPF_JSGT | BPF_X:
	case BPF_JMP32 | BPF_JSGT | BPF_K:

	case BPF_JMP | BPF_JSET | BPF_X:
	case BPF_JMP | BPF_JSET | BPF_K:
	case BPF_JMP32 | BPF_JSET | BPF_X:
	case BPF_JMP32 | BPF_JSET | BPF_K:
		rvoff = rv_offset(i, off, ctx);
		if (BPF_SRC(code) == BPF_K) {
			s = ctx->ninsns;
			emit_imm32(tmp2, imm, ctx);
			src = tmp2;
			e = ctx->ninsns;
			rvoff -= ninsns_rvoff(e - s);
		}

		if (is64)
			emit_branch_r64(dst, src, rvoff, ctx, BPF_OP(code));
		else
			emit_branch_r32(dst, src, rvoff, ctx, BPF_OP(code));
		break;

	case BPF_JMP | BPF_EXIT:
		if (i == ctx->prog->len - 1)
			break;

		rvoff = epilogue_offset(ctx);
		emit_jump_and_link(RV_REG_ZERO, rvoff, false, ctx);
		break;

	case BPF_LD | BPF_IMM | BPF_DW:
	{
		struct bpf_insn insn1 = insn[1];
		s32 imm_lo = imm;
		s32 imm_hi = insn1.imm;
		const s8 *rd = bpf_get_reg64(dst, tmp1, ctx);

		emit_imm64(rd, imm_hi, imm_lo, ctx);
		bpf_put_reg64(dst, rd, ctx);
		return 1;
	}

	case BPF_LDX | BPF_MEM | BPF_B:
	case BPF_LDX | BPF_MEM | BPF_H:
	case BPF_LDX | BPF_MEM | BPF_W:
	case BPF_LDX | BPF_MEM | BPF_DW:
		if (emit_load_r64(dst, src, off, ctx, BPF_SIZE(code)))
			return -1;
		break;

	case BPF_ST | BPF_MEM | BPF_B:
	case BPF_ST | BPF_MEM | BPF_H:
	case BPF_ST | BPF_MEM | BPF_W:
	case BPF_ST | BPF_MEM | BPF_DW:

	case BPF_STX | BPF_MEM | BPF_B:
	case BPF_STX | BPF_MEM | BPF_H:
	case BPF_STX | BPF_MEM | BPF_W:
	case BPF_STX | BPF_MEM | BPF_DW:
		if (BPF_CLASS(code) == BPF_ST) {
			emit_imm32(tmp2, imm, ctx);
			src = tmp2;
		}

		if (emit_store_r64(dst, src, off, ctx, BPF_SIZE(code),
				   BPF_MODE(code)))
			return -1;
		break;

	case BPF_STX | BPF_ATOMIC | BPF_W:
		if (insn->imm != BPF_ADD) {
			pr_info_once(
				"bpf-jit: not supported: atomic operation %02x ***\n",
				insn->imm);
			return -EFAULT;
		}

		if (emit_store_r64(dst, src, off, ctx, BPF_SIZE(code),
				   BPF_MODE(code)))
			return -1;
		break;

	/* No hardware support for 8-byte atomics in RV32. */
	case BPF_STX | BPF_ATOMIC | BPF_DW:
		/* Fallthrough. */

notsupported:
		pr_info_once("bpf-jit: not supported: opcode %02x ***\n", code);
		return -EFAULT;

	default:
		pr_err("bpf-jit: unknown opcode %02x\n", code);
		return -EINVAL;
	}

	return 0;
}

void bpf_jit_build_prologue(struct rv_jit_context *ctx)
{
	const s8 *fp = bpf2rv32[BPF_REG_FP];
	const s8 *r1 = bpf2rv32[BPF_REG_1];
	int stack_adjust = 0;
	int bpf_stack_adjust =
		round_up(ctx->prog->aux->stack_depth, STACK_ALIGN);

	/* Make space for callee-saved registers. */
	stack_adjust += NR_SAVED_REGISTERS * sizeof(u32);
	/* Make space for BPF registers on stack. */
	stack_adjust += BPF_JIT_SCRATCH_REGS * sizeof(u32);
	/* Make space for BPF stack. */
	stack_adjust += bpf_stack_adjust;
	/* Round up for stack alignment. */
	stack_adjust = round_up(stack_adjust, STACK_ALIGN);

	/*
	 * The first instruction sets the tail-call-counter (TCC) register.
	 * This instruction is skipped by tail calls.
	 */
	emit(rv_addi(RV_REG_TCC, RV_REG_ZERO, MAX_TAIL_CALL_CNT), ctx);

	emit(rv_addi(RV_REG_SP, RV_REG_SP, -stack_adjust), ctx);

	/* Save callee-save registers. */
	emit(rv_sw(RV_REG_SP, stack_adjust - 4, RV_REG_RA), ctx);
	emit(rv_sw(RV_REG_SP, stack_adjust - 8, RV_REG_FP), ctx);
	emit(rv_sw(RV_REG_SP, stack_adjust - 12, RV_REG_S1), ctx);
	emit(rv_sw(RV_REG_SP, stack_adjust - 16, RV_REG_S2), ctx);
	emit(rv_sw(RV_REG_SP, stack_adjust - 20, RV_REG_S3), ctx);
	emit(rv_sw(RV_REG_SP, stack_adjust - 24, RV_REG_S4), ctx);
	emit(rv_sw(RV_REG_SP, stack_adjust - 28, RV_REG_S5), ctx);
	emit(rv_sw(RV_REG_SP, stack_adjust - 32, RV_REG_S6), ctx);
	emit(rv_sw(RV_REG_SP, stack_adjust - 36, RV_REG_S7), ctx);

	/* Set fp: used as the base address for stacked BPF registers. */
	emit(rv_addi(RV_REG_FP, RV_REG_SP, stack_adjust), ctx);

	/* Set up BPF frame pointer. */
	emit(rv_addi(lo(fp), RV_REG_SP, bpf_stack_adjust), ctx);
	emit(rv_addi(hi(fp), RV_REG_ZERO, 0), ctx);

	/* Set up BPF context pointer. */
	emit(rv_addi(lo(r1), RV_REG_A0, 0), ctx);
	emit(rv_addi(hi(r1), RV_REG_ZERO, 0), ctx);

	ctx->stack_size = stack_adjust;
}

void bpf_jit_build_epilogue(struct rv_jit_context *ctx)
{
	__build_epilogue(false, ctx);
}