summaryrefslogtreecommitdiffstats
path: root/src/video_core/texture_cache/util.cpp
blob: 0a86ce13908855ed0c8b850e93d1d61c6e66ca5d (plain) (blame)
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
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-FileCopyrightText: Ryujinx Team and Contributors
// SPDX-License-Identifier: GPL-2.0-or-later AND MIT

// This files contains code from Ryujinx
// A copy of the code can be obtained from https://github.com/Ryujinx/Ryujinx
// The sections using code from Ryujinx are marked with a link to the original version

#include <algorithm>
#include <array>
#include <numeric>
#include <optional>
#include <span>
#include <vector>

#include "common/alignment.h"
#include "common/assert.h"
#include "common/bit_util.h"
#include "common/common_types.h"
#include "common/div_ceil.h"
#include "common/scratch_buffer.h"
#include "common/settings.h"
#include "core/memory.h"
#include "video_core/compatible_formats.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/memory_manager.h"
#include "video_core/surface.h"
#include "video_core/texture_cache/decode_bc.h"
#include "video_core/texture_cache/format_lookup_table.h"
#include "video_core/texture_cache/formatter.h"
#include "video_core/texture_cache/samples_helper.h"
#include "video_core/texture_cache/util.h"
#include "video_core/textures/astc.h"
#include "video_core/textures/bcn.h"
#include "video_core/textures/decoders.h"

namespace VideoCommon {

namespace {

using Tegra::Texture::GOB_SIZE;
using Tegra::Texture::GOB_SIZE_SHIFT;
using Tegra::Texture::GOB_SIZE_X;
using Tegra::Texture::GOB_SIZE_X_SHIFT;
using Tegra::Texture::GOB_SIZE_Y;
using Tegra::Texture::GOB_SIZE_Y_SHIFT;
using Tegra::Texture::GOB_SIZE_Z;
using Tegra::Texture::GOB_SIZE_Z_SHIFT;
using Tegra::Texture::MsaaMode;
using Tegra::Texture::SwizzleTexture;
using Tegra::Texture::TextureFormat;
using Tegra::Texture::TextureType;
using Tegra::Texture::TICEntry;
using Tegra::Texture::UnswizzleTexture;
using VideoCore::Surface::BytesPerBlock;
using VideoCore::Surface::DefaultBlockHeight;
using VideoCore::Surface::DefaultBlockWidth;
using VideoCore::Surface::IsCopyCompatible;
using VideoCore::Surface::IsPixelFormatASTC;
using VideoCore::Surface::IsViewCompatible;
using VideoCore::Surface::PixelFormatFromDepthFormat;
using VideoCore::Surface::PixelFormatFromRenderTargetFormat;
using VideoCore::Surface::SurfaceType;

struct LevelInfo {
    Extent3D size;
    Extent3D block;
    Extent2D tile_size;
    u32 bpp_log2;
    u32 tile_width_spacing;
};

[[nodiscard]] constexpr u32 AdjustTileSize(u32 shift, u32 unit_factor, u32 dimension) {
    if (shift == 0) {
        return 0;
    }
    u32 x = unit_factor << (shift - 1);
    if (x >= dimension) {
        while (--shift) {
            x >>= 1;
            if (x < dimension) {
                break;
            }
        }
    }
    return shift;
}

[[nodiscard]] constexpr u32 AdjustMipSize(u32 size, u32 level) {
    return std::max<u32>(size >> level, 1);
}

[[nodiscard]] constexpr Extent3D AdjustMipSize(Extent3D size, s32 level) {
    return Extent3D{
        .width = AdjustMipSize(size.width, level),
        .height = AdjustMipSize(size.height, level),
        .depth = AdjustMipSize(size.depth, level),
    };
}

[[nodiscard]] Extent3D AdjustSamplesSize(Extent3D size, s32 num_samples) {
    const auto [samples_x, samples_y] = SamplesLog2(num_samples);
    return Extent3D{
        .width = size.width >> samples_x,
        .height = size.height >> samples_y,
        .depth = size.depth,
    };
}

template <u32 GOB_EXTENT>
[[nodiscard]] constexpr u32 AdjustMipBlockSize(u32 num_tiles, u32 block_size, u32 level) {
    do {
        while (block_size > 0 && num_tiles <= (1U << (block_size - 1)) * GOB_EXTENT) {
            --block_size;
        }
    } while (level--);
    return block_size;
}

[[nodiscard]] constexpr Extent3D AdjustMipBlockSize(Extent3D num_tiles, Extent3D block_size,
                                                    u32 level) {
    return {
        .width = AdjustMipBlockSize<GOB_SIZE_X>(num_tiles.width, block_size.width, level),
        .height = AdjustMipBlockSize<GOB_SIZE_Y>(num_tiles.height, block_size.height, level),
        .depth = level == 0
                     ? block_size.depth
                     : AdjustMipBlockSize<GOB_SIZE_Z>(num_tiles.depth, block_size.depth, level),
    };
}

[[nodiscard]] constexpr Extent3D AdjustTileSize(Extent3D size, Extent2D tile_size) {
    return {
        .width = Common::DivCeil(size.width, tile_size.width),
        .height = Common::DivCeil(size.height, tile_size.height),
        .depth = size.depth,
    };
}

[[nodiscard]] constexpr u32 BytesPerBlockLog2(u32 bytes_per_block) {
    return std::countl_zero(bytes_per_block) ^ 0x1F;
}

[[nodiscard]] constexpr u32 BytesPerBlockLog2(PixelFormat format) {
    return BytesPerBlockLog2(BytesPerBlock(format));
}

[[nodiscard]] constexpr u32 NumBlocks(Extent3D size, Extent2D tile_size) {
    const Extent3D num_blocks = AdjustTileSize(size, tile_size);
    return num_blocks.width * num_blocks.height * num_blocks.depth;
}

[[nodiscard]] constexpr u32 AdjustSize(u32 size, u32 level, u32 block_size) {
    return Common::DivCeil(AdjustMipSize(size, level), block_size);
}

[[nodiscard]] constexpr Extent2D DefaultBlockSize(PixelFormat format) {
    return {DefaultBlockWidth(format), DefaultBlockHeight(format)};
}

[[nodiscard]] constexpr Extent3D NumLevelBlocks(const LevelInfo& info, u32 level) {
    return Extent3D{
        .width = AdjustSize(info.size.width, level, info.tile_size.width) << info.bpp_log2,
        .height = AdjustSize(info.size.height, level, info.tile_size.height),
        .depth = AdjustMipSize(info.size.depth, level),
    };
}

[[nodiscard]] constexpr Extent3D TileShift(const LevelInfo& info, u32 level) {
    if (level == 0) {
        return Extent3D{
            .width = info.block.width,
            .height = info.block.height,
            .depth = info.block.depth,
        };
    }
    const Extent3D blocks = NumLevelBlocks(info, level);
    return Extent3D{
        .width = AdjustTileSize(info.block.width, GOB_SIZE_X, blocks.width),
        .height = AdjustTileSize(info.block.height, GOB_SIZE_Y, blocks.height),
        .depth = AdjustTileSize(info.block.depth, GOB_SIZE_Z, blocks.depth),
    };
}

[[nodiscard]] constexpr Extent2D GobSize(u32 bpp_log2, u32 block_height, u32 tile_width_spacing) {
    return Extent2D{
        .width = GOB_SIZE_X_SHIFT - bpp_log2 + tile_width_spacing,
        .height = GOB_SIZE_Y_SHIFT + block_height,
    };
}

[[nodiscard]] constexpr bool IsSmallerThanGobSize(Extent3D num_tiles, Extent2D gob,
                                                  u32 block_depth) {
    return num_tiles.width <= (1U << gob.width) || num_tiles.height <= (1U << gob.height) ||
           num_tiles.depth < (1U << block_depth);
}

[[nodiscard]] constexpr u32 StrideAlignment(Extent3D num_tiles, Extent3D block, Extent2D gob,
                                            u32 bpp_log2) {
    if (IsSmallerThanGobSize(num_tiles, gob, block.depth)) {
        return GOB_SIZE_X_SHIFT - bpp_log2;
    } else {
        return gob.width;
    }
}

[[nodiscard]] constexpr u32 StrideAlignment(Extent3D num_tiles, Extent3D block, u32 bpp_log2,
                                            u32 tile_width_spacing) {
    const Extent2D gob = GobSize(bpp_log2, block.height, tile_width_spacing);
    return StrideAlignment(num_tiles, block, gob, bpp_log2);
}

[[nodiscard]] constexpr Extent2D NumGobs(const LevelInfo& info, u32 level) {
    const Extent3D blocks = NumLevelBlocks(info, level);
    const Extent2D gobs{
        .width = Common::DivCeilLog2(blocks.width, GOB_SIZE_X_SHIFT),
        .height = Common::DivCeilLog2(blocks.height, GOB_SIZE_Y_SHIFT),
    };
    const Extent2D gob = GobSize(info.bpp_log2, info.block.height, info.tile_width_spacing);
    const bool is_small = IsSmallerThanGobSize(blocks, gob, info.block.depth);
    const u32 alignment = is_small ? 0 : info.tile_width_spacing;
    return Extent2D{
        .width = Common::AlignUpLog2(gobs.width, alignment),
        .height = gobs.height,
    };
}

[[nodiscard]] constexpr Extent3D LevelTiles(const LevelInfo& info, u32 level) {
    const Extent3D blocks = NumLevelBlocks(info, level);
    const Extent3D tile_shift = TileShift(info, level);
    const Extent2D gobs = NumGobs(info, level);
    return Extent3D{
        .width = Common::DivCeilLog2(gobs.width, tile_shift.width),
        .height = Common::DivCeilLog2(gobs.height, tile_shift.height),
        .depth = Common::DivCeilLog2(blocks.depth, tile_shift.depth),
    };
}

[[nodiscard]] constexpr u32 CalculateLevelSize(const LevelInfo& info, u32 level) {
    const Extent3D tile_shift = TileShift(info, level);
    const Extent3D tiles = LevelTiles(info, level);
    const u32 num_tiles = tiles.width * tiles.height * tiles.depth;
    const u32 shift = GOB_SIZE_SHIFT + tile_shift.width + tile_shift.height + tile_shift.depth;
    return num_tiles << shift;
}

[[nodiscard]] constexpr LevelArray CalculateLevelSizes(const LevelInfo& info, u32 num_levels) {
    ASSERT(num_levels <= MAX_MIP_LEVELS);
    LevelArray sizes{};
    for (u32 level = 0; level < num_levels; ++level) {
        sizes[level] = CalculateLevelSize(info, level);
    }
    return sizes;
}

[[nodiscard]] u32 CalculateLevelBytes(const LevelArray& sizes, u32 num_levels) {
    return std::reduce(sizes.begin(), sizes.begin() + num_levels, 0U);
}

[[nodiscard]] constexpr LevelInfo MakeLevelInfo(PixelFormat format, Extent3D size, Extent3D block,
                                                u32 tile_width_spacing) {
    const u32 bytes_per_block = BytesPerBlock(format);
    return {
        .size =
            {
                .width = size.width,
                .height = size.height,
                .depth = size.depth,
            },
        .block = block,
        .tile_size = DefaultBlockSize(format),
        .bpp_log2 = BytesPerBlockLog2(bytes_per_block),
        .tile_width_spacing = tile_width_spacing,
    };
}

[[nodiscard]] constexpr LevelInfo MakeLevelInfo(const ImageInfo& info) {
    return MakeLevelInfo(info.format, info.size, info.block, info.tile_width_spacing);
}

[[nodiscard]] constexpr u32 CalculateLevelOffset(PixelFormat format, Extent3D size, Extent3D block,
                                                 u32 tile_width_spacing, u32 level) {
    const LevelInfo info = MakeLevelInfo(format, size, block, tile_width_spacing);
    u32 offset = 0;
    for (u32 current_level = 0; current_level < level; ++current_level) {
        offset += CalculateLevelSize(info, current_level);
    }
    return offset;
}

[[nodiscard]] constexpr u32 AlignLayerSize(u32 size_bytes, Extent3D size, Extent3D block,
                                           u32 tile_size_y, u32 tile_width_spacing) {
    // https://github.com/Ryujinx/Ryujinx/blob/1c9aba6de1520aea5480c032e0ff5664ac1bb36f/Ryujinx.Graphics.Texture/SizeCalculator.cs#L134
    if (tile_width_spacing > 0) {
        const u32 alignment_log2 = GOB_SIZE_SHIFT + tile_width_spacing + block.height + block.depth;
        return Common::AlignUpLog2(size_bytes, alignment_log2);
    }
    const u32 aligned_height = Common::AlignUp(size.height, tile_size_y);
    while (block.height != 0 && aligned_height <= (1U << (block.height - 1)) * GOB_SIZE_Y) {
        --block.height;
    }
    while (block.depth != 0 && size.depth <= (1U << (block.depth - 1))) {
        --block.depth;
    }
    const u32 block_shift = GOB_SIZE_SHIFT + block.height + block.depth;
    const u32 num_blocks = size_bytes >> block_shift;
    if (size_bytes != num_blocks << block_shift) {
        return (num_blocks + 1) << block_shift;
    }
    return size_bytes;
}

[[nodiscard]] std::optional<SubresourceExtent> ResolveOverlapEqualAddress(const ImageInfo& new_info,
                                                                          const ImageBase& overlap,
                                                                          bool strict_size) {
    const ImageInfo& info = overlap.info;
    if (!IsBlockLinearSizeCompatible(new_info, info, 0, 0, strict_size)) {
        return std::nullopt;
    }
    if (new_info.block != info.block) {
        return std::nullopt;
    }
    const SubresourceExtent resources = new_info.resources;
    return SubresourceExtent{
        .levels = std::max(resources.levels, info.resources.levels),
        .layers = std::max(resources.layers, info.resources.layers),
    };
}

[[nodiscard]] std::optional<SubresourceExtent> ResolveOverlapRightAddress3D(
    const ImageInfo& new_info, GPUVAddr gpu_addr, const ImageBase& overlap, bool strict_size) {
    const auto slice_offsets = CalculateSliceOffsets(new_info);
    const u32 diff = static_cast<u32>(overlap.gpu_addr - gpu_addr);
    const auto it = std::ranges::find(slice_offsets, diff);
    if (it == slice_offsets.end()) {
        return std::nullopt;
    }
    const auto subresources = CalculateSliceSubresources(new_info);
    const SubresourceBase base = subresources[std::distance(slice_offsets.begin(), it)];
    const ImageInfo& info = overlap.info;
    if (!IsBlockLinearSizeCompatible(new_info, info, base.level, 0, strict_size)) {
        return std::nullopt;
    }
    const u32 mip_depth = std::max(1U, new_info.size.depth << base.level);
    if (mip_depth < info.size.depth + base.layer) {
        return std::nullopt;
    }
    if (MipBlockSize(new_info, base.level) != info.block) {
        return std::nullopt;
    }
    return SubresourceExtent{
        .levels = std::max(new_info.resources.levels, info.resources.levels + base.level),
        .layers = 1,
    };
}

[[nodiscard]] std::optional<SubresourceExtent> ResolveOverlapRightAddress2D(
    const ImageInfo& new_info, GPUVAddr gpu_addr, const ImageBase& overlap, bool strict_size) {
    const u64 layer_stride = new_info.layer_stride;
    const u64 new_size = layer_stride * new_info.resources.layers;
    const u64 diff = overlap.gpu_addr - gpu_addr;
    if (diff > new_size) {
        return std::nullopt;
    }
    const s32 base_layer = static_cast<s32>(diff / layer_stride);
    const s32 mip_offset = static_cast<s32>(diff % layer_stride);
    const std::array offsets = CalculateMipLevelOffsets(new_info);
    const auto end = offsets.begin() + new_info.resources.levels;
    const auto it = std::find(offsets.begin(), end, static_cast<u32>(mip_offset));
    if (it == end) {
        // Mipmap is not aligned to any valid size
        return std::nullopt;
    }
    const SubresourceBase base{
        .level = static_cast<s32>(std::distance(offsets.begin(), it)),
        .layer = base_layer,
    };
    const ImageInfo& info = overlap.info;
    if (!IsBlockLinearSizeCompatible(new_info, info, base.level, 0, strict_size)) {
        return std::nullopt;
    }
    if (MipBlockSize(new_info, base.level) != info.block) {
        return std::nullopt;
    }
    return SubresourceExtent{
        .levels = std::max(new_info.resources.levels, info.resources.levels + base.level),
        .layers = std::max(new_info.resources.layers, info.resources.layers + base.layer),
    };
}

[[nodiscard]] std::optional<OverlapResult> ResolveOverlapRightAddress(const ImageInfo& new_info,
                                                                      GPUVAddr gpu_addr,
                                                                      VAddr cpu_addr,
                                                                      const ImageBase& overlap,
                                                                      bool strict_size) {
    std::optional<SubresourceExtent> resources;
    if (new_info.type != ImageType::e3D) {
        resources = ResolveOverlapRightAddress2D(new_info, gpu_addr, overlap, strict_size);
    } else {
        resources = ResolveOverlapRightAddress3D(new_info, gpu_addr, overlap, strict_size);
    }
    if (!resources) {
        return std::nullopt;
    }
    return OverlapResult{
        .gpu_addr = gpu_addr,
        .cpu_addr = cpu_addr,
        .resources = *resources,
    };
}

[[nodiscard]] std::optional<OverlapResult> ResolveOverlapLeftAddress(const ImageInfo& new_info,
                                                                     GPUVAddr gpu_addr,
                                                                     VAddr cpu_addr,
                                                                     const ImageBase& overlap,
                                                                     bool strict_size) {
    const std::optional<SubresourceBase> base = overlap.TryFindBase(gpu_addr);
    if (!base) {
        return std::nullopt;
    }
    const ImageInfo& info = overlap.info;
    if (!IsBlockLinearSizeCompatible(new_info, info, base->level, 0, strict_size)) {
        return std::nullopt;
    }
    if (new_info.block != MipBlockSize(info, base->level)) {
        return std::nullopt;
    }
    const SubresourceExtent resources = new_info.resources;
    s32 layers = 1;
    if (info.type != ImageType::e3D) {
        layers = std::max(resources.layers, info.resources.layers + base->layer);
    }
    return OverlapResult{
        .gpu_addr = overlap.gpu_addr,
        .cpu_addr = overlap.cpu_addr,
        .resources =
            {
                .levels = std::max(resources.levels + base->level, info.resources.levels),
                .layers = layers,
            },
    };
}

[[nodiscard]] Extent2D PitchLinearAlignedSize(const ImageInfo& info) {
    // https://github.com/Ryujinx/Ryujinx/blob/1c9aba6de1520aea5480c032e0ff5664ac1bb36f/Ryujinx.Graphics.Texture/SizeCalculator.cs#L212
    static constexpr u32 STRIDE_ALIGNMENT = 32;
    ASSERT(info.type == ImageType::Linear);
    const Extent2D num_tiles{
        .width = Common::DivCeil(info.size.width, DefaultBlockWidth(info.format)),
        .height = Common::DivCeil(info.size.height, DefaultBlockHeight(info.format)),
    };
    const u32 width_alignment = STRIDE_ALIGNMENT / BytesPerBlock(info.format);
    return Extent2D{
        .width = Common::AlignUp(num_tiles.width, width_alignment),
        .height = num_tiles.height,
    };
}

[[nodiscard]] Extent3D BlockLinearAlignedSize(const ImageInfo& info, u32 level) {
    // https://github.com/Ryujinx/Ryujinx/blob/1c9aba6de1520aea5480c032e0ff5664ac1bb36f/Ryujinx.Graphics.Texture/SizeCalculator.cs#L176
    ASSERT(info.type != ImageType::Linear);
    const Extent3D size = AdjustMipSize(info.size, level);
    const Extent3D num_tiles{
        .width = Common::DivCeil(size.width, DefaultBlockWidth(info.format)),
        .height = Common::DivCeil(size.height, DefaultBlockHeight(info.format)),
        .depth = size.depth,
    };
    const u32 bpp_log2 = BytesPerBlockLog2(info.format);
    const u32 alignment = StrideAlignment(num_tiles, info.block, bpp_log2, info.tile_width_spacing);
    const Extent3D mip_block = AdjustMipBlockSize(num_tiles, info.block, 0);
    return Extent3D{
        .width = Common::AlignUpLog2(num_tiles.width, alignment),
        .height = Common::AlignUpLog2(num_tiles.height, GOB_SIZE_Y_SHIFT + mip_block.height),
        .depth = Common::AlignUpLog2(num_tiles.depth, GOB_SIZE_Z_SHIFT + mip_block.depth),
    };
}

[[nodiscard]] constexpr u32 NumBlocksPerLayer(const ImageInfo& info, Extent2D tile_size) noexcept {
    u32 num_blocks = 0;
    for (s32 level = 0; level < info.resources.levels; ++level) {
        const Extent3D mip_size = AdjustMipSize(info.size, level);
        num_blocks += NumBlocks(mip_size, tile_size);
    }
    return num_blocks;
}

[[nodiscard]] u32 NumSlices(const ImageInfo& info) noexcept {
    ASSERT(info.type == ImageType::e3D);
    u32 num_slices = 0;
    for (s32 level = 0; level < info.resources.levels; ++level) {
        num_slices += AdjustMipSize(info.size.depth, level);
    }
    return num_slices;
}

void SwizzlePitchLinearImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr,
                             const ImageInfo& info, const BufferImageCopy& copy,
                             std::span<const u8> memory) {
    ASSERT(copy.image_offset.z == 0);
    ASSERT(copy.image_extent.depth == 1);
    ASSERT(copy.image_subresource.base_level == 0);
    ASSERT(copy.image_subresource.base_layer == 0);
    ASSERT(copy.image_subresource.num_layers == 1);

    const u32 bytes_per_block = BytesPerBlock(info.format);
    const u32 row_length = copy.image_extent.width * bytes_per_block;
    const u32 guest_offset_x = copy.image_offset.x * bytes_per_block;

    for (u32 line = 0; line < copy.image_extent.height; ++line) {
        const u32 host_offset_y = line * info.pitch;
        const u32 guest_offset_y = (copy.image_offset.y + line) * info.pitch;
        const u32 guest_offset = guest_offset_x + guest_offset_y;
        gpu_memory.WriteBlockUnsafe(gpu_addr + guest_offset, memory.data() + host_offset_y,
                                    row_length);
    }
}

void SwizzleBlockLinearImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr,
                             const ImageInfo& info, const BufferImageCopy& copy,
                             std::span<const u8> input, Common::ScratchBuffer<u8>& tmp_buffer) {
    const Extent3D size = info.size;
    const LevelInfo level_info = MakeLevelInfo(info);
    const Extent2D tile_size = DefaultBlockSize(info.format);
    const u32 bytes_per_block = BytesPerBlock(info.format);

    const s32 level = copy.image_subresource.base_level;
    const Extent3D level_size = AdjustMipSize(size, level);
    const u32 num_blocks_per_layer = NumBlocks(level_size, tile_size);
    const u32 host_bytes_per_layer = num_blocks_per_layer * bytes_per_block;

    UNIMPLEMENTED_IF(copy.image_offset.x != 0);
    UNIMPLEMENTED_IF(copy.image_offset.y != 0);
    UNIMPLEMENTED_IF(copy.image_offset.z != 0);
    UNIMPLEMENTED_IF(copy.image_extent != level_size);

    const Extent3D num_tiles = AdjustTileSize(level_size, tile_size);
    const Extent3D block = AdjustMipBlockSize(num_tiles, level_info.block, level);

    size_t host_offset = copy.buffer_offset;

    const u32 num_levels = info.resources.levels;
    const std::array sizes = CalculateLevelSizes(level_info, num_levels);
    size_t guest_offset = CalculateLevelBytes(sizes, level);
    const size_t layer_stride =
        AlignLayerSize(CalculateLevelBytes(sizes, num_levels), size, level_info.block,
                       tile_size.height, info.tile_width_spacing);
    const size_t subresource_size = sizes[level];

    for (s32 layer = 0; layer < info.resources.layers; ++layer) {
        const std::span<const u8> src = input.subspan(host_offset);
        {
            Core::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::UnsafeReadWrite>
                dst(gpu_memory, gpu_addr + guest_offset, subresource_size, &tmp_buffer);

            SwizzleTexture(dst, src, bytes_per_block, num_tiles.width, num_tiles.height,
                           num_tiles.depth, block.height, block.depth);
        }

        host_offset += host_bytes_per_layer;
        guest_offset += layer_stride;
    }
    ASSERT(host_offset - copy.buffer_offset == copy.buffer_size);
}

} // Anonymous namespace

u32 CalculateGuestSizeInBytes(const ImageInfo& info) noexcept {
    if (info.type == ImageType::Buffer) {
        return info.size.width * BytesPerBlock(info.format);
    }
    if (info.type == ImageType::Linear) {
        return info.pitch * Common::DivCeil(info.size.height, DefaultBlockHeight(info.format));
    }
    if (info.resources.layers > 1) {
        ASSERT(info.layer_stride != 0);
        return info.layer_stride * info.resources.layers;
    } else {
        return CalculateLayerSize(info);
    }
}

u32 CalculateUnswizzledSizeBytes(const ImageInfo& info) noexcept {
    if (info.type == ImageType::Buffer) {
        return info.size.width * BytesPerBlock(info.format);
    }
    if (info.type == ImageType::Linear) {
        return info.pitch * Common::DivCeil(info.size.height, DefaultBlockHeight(info.format));
    }
    const Extent2D tile_size = DefaultBlockSize(info.format);
    return NumBlocksPerLayer(info, tile_size) * info.resources.layers * BytesPerBlock(info.format);
}

u32 CalculateConvertedSizeBytes(const ImageInfo& info) noexcept {
    if (info.type == ImageType::Buffer) {
        return info.size.width * BytesPerBlock(info.format);
    }
    static constexpr Extent2D TILE_SIZE{1, 1};
    if (IsPixelFormatASTC(info.format) && Settings::values.astc_recompression.GetValue() !=
                                              Settings::AstcRecompression::Uncompressed) {
        const u32 bpp_div =
            Settings::values.astc_recompression.GetValue() == Settings::AstcRecompression::Bc1 ? 2
                                                                                               : 1;
        // NumBlocksPerLayer doesn't account for this correctly, so we have to do it manually.
        u32 output_size = 0;
        for (s32 i = 0; i < info.resources.levels; i++) {
            const auto mip_size = AdjustMipSize(info.size, i);
            const u32 plane_dim =
                Common::AlignUp(mip_size.width, 4U) * Common::AlignUp(mip_size.height, 4U);
            output_size += (plane_dim * info.size.depth * info.resources.layers) / bpp_div;
        }
        return output_size;
    }
    return NumBlocksPerLayer(info, TILE_SIZE) * info.resources.layers *
           ConvertedBytesPerBlock(info.format);
}

u32 CalculateLayerStride(const ImageInfo& info) noexcept {
    ASSERT(info.type != ImageType::Linear);
    const u32 layer_size = CalculateLayerSize(info);
    const Extent3D size = info.size;
    const Extent3D block = info.block;
    const u32 tile_size_y = DefaultBlockHeight(info.format);
    return AlignLayerSize(layer_size, size, block, tile_size_y, info.tile_width_spacing);
}

u32 CalculateLayerSize(const ImageInfo& info) noexcept {
    ASSERT(info.type != ImageType::Linear);
    return CalculateLevelOffset(info.format, info.size, info.block, info.tile_width_spacing,
                                info.resources.levels);
}

LevelArray CalculateMipLevelOffsets(const ImageInfo& info) noexcept {
    if (info.type == ImageType::Linear) {
        return {};
    }
    ASSERT(info.resources.levels <= static_cast<s32>(MAX_MIP_LEVELS));
    const LevelInfo level_info = MakeLevelInfo(info);
    LevelArray offsets{};
    u32 offset = 0;
    for (s32 level = 0; level < info.resources.levels; ++level) {
        offsets[level] = offset;
        offset += CalculateLevelSize(level_info, level);
    }
    return offsets;
}

LevelArray CalculateMipLevelSizes(const ImageInfo& info) noexcept {
    const u32 num_levels = info.resources.levels;
    const LevelInfo level_info = MakeLevelInfo(info);
    LevelArray sizes{};
    for (u32 level = 0; level < num_levels; ++level) {
        sizes[level] = CalculateLevelSize(level_info, level);
    }
    return sizes;
}

boost::container::small_vector<u32, 16> CalculateSliceOffsets(const ImageInfo& info) {
    ASSERT(info.type == ImageType::e3D);
    boost::container::small_vector<u32, 16> offsets;
    offsets.reserve(NumSlices(info));

    const LevelInfo level_info = MakeLevelInfo(info);
    u32 mip_offset = 0;
    for (s32 level = 0; level < info.resources.levels; ++level) {
        const Extent3D tile_shift = TileShift(level_info, level);
        const Extent3D tiles = LevelTiles(level_info, level);
        const u32 gob_size_shift = tile_shift.height + GOB_SIZE_SHIFT;
        const u32 slice_size = (tiles.width * tiles.height) << gob_size_shift;
        const u32 z_mask = (1U << tile_shift.depth) - 1;
        const u32 depth = AdjustMipSize(info.size.depth, level);
        for (u32 slice = 0; slice < depth; ++slice) {
            const u32 z_low = slice & z_mask;
            const u32 z_high = slice & ~z_mask;
            offsets.push_back(mip_offset + (z_low << gob_size_shift) + (z_high * slice_size));
        }
        mip_offset += CalculateLevelSize(level_info, level);
    }
    return offsets;
}

boost::container::small_vector<SubresourceBase, 16> CalculateSliceSubresources(
    const ImageInfo& info) {
    ASSERT(info.type == ImageType::e3D);
    boost::container::small_vector<SubresourceBase, 16> subresources;
    subresources.reserve(NumSlices(info));
    for (s32 level = 0; level < info.resources.levels; ++level) {
        const s32 depth = AdjustMipSize(info.size.depth, level);
        for (s32 slice = 0; slice < depth; ++slice) {
            subresources.emplace_back(SubresourceBase{
                .level = level,
                .layer = slice,
            });
        }
    }
    return subresources;
}

u32 CalculateLevelStrideAlignment(const ImageInfo& info, u32 level) {
    const Extent2D tile_size = DefaultBlockSize(info.format);
    const Extent3D level_size = AdjustMipSize(info.size, level);
    const Extent3D num_tiles = AdjustTileSize(level_size, tile_size);
    const Extent3D block = AdjustMipBlockSize(num_tiles, info.block, level);
    const u32 bpp_log2 = BytesPerBlockLog2(info.format);
    return StrideAlignment(num_tiles, block, bpp_log2, info.tile_width_spacing);
}

PixelFormat PixelFormatFromTIC(const TICEntry& config) noexcept {
    return PixelFormatFromTextureInfo(config.format, config.r_type, config.g_type, config.b_type,
                                      config.a_type, config.srgb_conversion);
}

ImageViewType RenderTargetImageViewType(const ImageInfo& info) noexcept {
    switch (info.type) {
    case ImageType::e2D:
        return info.resources.layers > 1 ? ImageViewType::e2DArray : ImageViewType::e2D;
    case ImageType::e3D:
        return ImageViewType::e2DArray;
    case ImageType::Linear:
        return ImageViewType::e2D;
    default:
        UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<int>(info.type));
        return ImageViewType{};
    }
}

boost::container::small_vector<ImageCopy, 16> MakeShrinkImageCopies(const ImageInfo& dst,
                                                                    const ImageInfo& src,
                                                                    SubresourceBase base,
                                                                    u32 up_scale, u32 down_shift) {
    ASSERT(dst.resources.levels >= src.resources.levels);

    const bool is_dst_3d = dst.type == ImageType::e3D;
    if (is_dst_3d) {
        ASSERT(src.type == ImageType::e3D);
        ASSERT(src.resources.levels == 1);
    }
    const bool both_2d{src.type == ImageType::e2D && dst.type == ImageType::e2D};
    boost::container::small_vector<ImageCopy, 16> copies;
    copies.reserve(src.resources.levels);
    for (s32 level = 0; level < src.resources.levels; ++level) {
        ImageCopy& copy = copies.emplace_back();
        copy.src_subresource = SubresourceLayers{
            .base_level = level,
            .base_layer = 0,
            .num_layers = src.resources.layers,
        };
        copy.dst_subresource = SubresourceLayers{
            .base_level = base.level + level,
            .base_layer = is_dst_3d ? 0 : base.layer,
            .num_layers = is_dst_3d ? 1 : src.resources.layers,
        };
        copy.src_offset = Offset3D{
            .x = 0,
            .y = 0,
            .z = 0,
        };
        copy.dst_offset = Offset3D{
            .x = 0,
            .y = 0,
            .z = is_dst_3d ? base.layer : 0,
        };
        const Extent3D mip_size = AdjustMipSize(dst.size, base.level + level);
        copy.extent = AdjustSamplesSize(mip_size, dst.num_samples);
        if (is_dst_3d) {
            copy.extent.depth = src.size.depth;
        }
        copy.extent.width = std::max<u32>((copy.extent.width * up_scale) >> down_shift, 1);
        if (both_2d) {
            copy.extent.height = std::max<u32>((copy.extent.height * up_scale) >> down_shift, 1);
        }
    }
    return copies;
}

boost::container::small_vector<ImageCopy, 16> MakeReinterpretImageCopies(const ImageInfo& src,
                                                                         u32 up_scale,
                                                                         u32 down_shift) {
    boost::container::small_vector<ImageCopy, 16> copies;
    copies.reserve(src.resources.levels);
    const bool is_3d = src.type == ImageType::e3D;
    for (s32 level = 0; level < src.resources.levels; ++level) {
        ImageCopy& copy = copies.emplace_back();
        copy.src_subresource = SubresourceLayers{
            .base_level = level,
            .base_layer = 0,
            .num_layers = src.resources.layers,
        };
        copy.dst_subresource = SubresourceLayers{
            .base_level = level,
            .base_layer = 0,
            .num_layers = src.resources.layers,
        };
        copy.src_offset = Offset3D{
            .x = 0,
            .y = 0,
            .z = 0,
        };
        copy.dst_offset = Offset3D{
            .x = 0,
            .y = 0,
            .z = 0,
        };
        const Extent3D mip_size = AdjustMipSize(src.size, level);
        copy.extent = AdjustSamplesSize(mip_size, src.num_samples);
        if (is_3d) {
            copy.extent.depth = src.size.depth;
        }
        copy.extent.width = std::max<u32>((copy.extent.width * up_scale) >> down_shift, 1);
        copy.extent.height = std::max<u32>((copy.extent.height * up_scale) >> down_shift, 1);
    }
    return copies;
}

bool IsValidEntry(const Tegra::MemoryManager& gpu_memory, const TICEntry& config) {
    const GPUVAddr address = config.Address();
    if (address == 0) {
        return false;
    }
    if (address >= (1ULL << 40)) {
        return false;
    }
    if (gpu_memory.GpuToCpuAddress(address).has_value()) {
        return true;
    }
    const ImageInfo info{config};
    const size_t guest_size_bytes = CalculateGuestSizeInBytes(info);
    return gpu_memory.GpuToCpuAddress(address, guest_size_bytes).has_value();
}

boost::container::small_vector<BufferImageCopy, 16> UnswizzleImage(Tegra::MemoryManager& gpu_memory,
                                                                   GPUVAddr gpu_addr,
                                                                   const ImageInfo& info,
                                                                   std::span<const u8> input,
                                                                   std::span<u8> output) {
    const size_t guest_size_bytes = input.size_bytes();
    const u32 bpp_log2 = BytesPerBlockLog2(info.format);
    const Extent3D size = info.size;

    if (info.type == ImageType::Linear) {
        ASSERT(output.size_bytes() >= guest_size_bytes);
        gpu_memory.ReadBlockUnsafe(gpu_addr, output.data(), guest_size_bytes);

        ASSERT((info.pitch >> bpp_log2) << bpp_log2 == info.pitch);
        return {{
            .buffer_offset = 0,
            .buffer_size = guest_size_bytes,
            .buffer_row_length = info.pitch >> bpp_log2,
            .buffer_image_height = size.height,
            .image_subresource =
                {
                    .base_level = 0,
                    .base_layer = 0,
                    .num_layers = 1,
                },
            .image_offset = {0, 0, 0},
            .image_extent = size,
        }};
    }
    const LevelInfo level_info = MakeLevelInfo(info);
    const s32 num_layers = info.resources.layers;
    const s32 num_levels = info.resources.levels;
    const Extent2D tile_size = DefaultBlockSize(info.format);
    const std::array level_sizes = CalculateLevelSizes(level_info, num_levels);
    const Extent2D gob = GobSize(bpp_log2, info.block.height, info.tile_width_spacing);
    const u32 layer_size = CalculateLevelBytes(level_sizes, num_levels);
    const u32 layer_stride = AlignLayerSize(layer_size, size, level_info.block, tile_size.height,
                                            info.tile_width_spacing);
    size_t guest_offset = 0;
    u32 host_offset = 0;
    boost::container::small_vector<BufferImageCopy, 16> copies(num_levels);

    for (s32 level = 0; level < num_levels; ++level) {
        const Extent3D level_size = AdjustMipSize(size, level);
        const u32 num_blocks_per_layer = NumBlocks(level_size, tile_size);
        const u32 host_bytes_per_layer = num_blocks_per_layer << bpp_log2;
        copies[level] = BufferImageCopy{
            .buffer_offset = host_offset,
            .buffer_size = static_cast<size_t>(host_bytes_per_layer) * num_layers,
            .buffer_row_length = Common::AlignUp(level_size.width, tile_size.width),
            .buffer_image_height = Common::AlignUp(level_size.height, tile_size.height),
            .image_subresource =
                {
                    .base_level = level,
                    .base_layer = 0,
                    .num_layers = info.resources.layers,
                },
            .image_offset = {0, 0, 0},
            .image_extent = level_size,
        };
        const Extent3D num_tiles = AdjustTileSize(level_size, tile_size);
        const Extent3D block = AdjustMipBlockSize(num_tiles, level_info.block, level);
        const u32 stride_alignment = StrideAlignment(num_tiles, info.block, gob, bpp_log2);
        size_t guest_layer_offset = 0;

        for (s32 layer = 0; layer < info.resources.layers; ++layer) {
            const std::span<u8> dst = output.subspan(host_offset);
            const std::span<const u8> src = input.subspan(guest_offset + guest_layer_offset);
            UnswizzleTexture(dst, src, 1U << bpp_log2, num_tiles.width, num_tiles.height,
                             num_tiles.depth, block.height, block.depth, stride_alignment);
            guest_layer_offset += layer_stride;
            host_offset += host_bytes_per_layer;
        }
        guest_offset += level_sizes[level];
    }
    return copies;
}

void ConvertImage(std::span<const u8> input, const ImageInfo& info, std::span<u8> output,
                  std::span<BufferImageCopy> copies) {
    u32 output_offset = 0;
    Common::ScratchBuffer<u8> decode_scratch;

    const Extent2D tile_size = DefaultBlockSize(info.format);
    for (BufferImageCopy& copy : copies) {
        const u32 level = copy.image_subresource.base_level;
        const Extent3D mip_size = AdjustMipSize(info.size, level);
        ASSERT(copy.image_offset == Offset3D{});
        ASSERT(copy.image_subresource.base_layer == 0);
        ASSERT(copy.image_extent == mip_size);
        ASSERT(copy.buffer_row_length == Common::AlignUp(mip_size.width, tile_size.width));
        ASSERT(copy.buffer_image_height == Common::AlignUp(mip_size.height, tile_size.height));

        const auto input_offset = input.subspan(copy.buffer_offset);
        copy.buffer_offset = output_offset;
        copy.buffer_row_length = mip_size.width;
        copy.buffer_image_height = mip_size.height;

        const auto recompression_setting = Settings::values.astc_recompression.GetValue();
        const bool astc = IsPixelFormatASTC(info.format);

        if (astc && recompression_setting == Settings::AstcRecompression::Uncompressed) {
            Tegra::Texture::ASTC::Decompress(
                input_offset, copy.image_extent.width, copy.image_extent.height,
                copy.image_subresource.num_layers * copy.image_extent.depth, tile_size.width,
                tile_size.height, output.subspan(output_offset));

            output_offset += copy.image_extent.width * copy.image_extent.height *
                             copy.image_subresource.num_layers *
                             BytesPerBlock(PixelFormat::A8B8G8R8_UNORM);
        } else if (astc) {
            // BC1 uses 0.5 bytes per texel
            // BC3 uses 1 byte per texel
            const auto compress = recompression_setting == Settings::AstcRecompression::Bc1
                                      ? Tegra::Texture::BCN::CompressBC1
                                      : Tegra::Texture::BCN::CompressBC3;
            const auto bpp_div = recompression_setting == Settings::AstcRecompression::Bc1 ? 2 : 1;

            const u32 plane_dim = copy.image_extent.width * copy.image_extent.height;
            const u32 level_size = plane_dim * copy.image_extent.depth *
                                   copy.image_subresource.num_layers *
                                   BytesPerBlock(PixelFormat::A8B8G8R8_UNORM);
            decode_scratch.resize_destructive(level_size);

            Tegra::Texture::ASTC::Decompress(
                input_offset, copy.image_extent.width, copy.image_extent.height,
                copy.image_subresource.num_layers * copy.image_extent.depth, tile_size.width,
                tile_size.height, decode_scratch);

            compress(decode_scratch, copy.image_extent.width, copy.image_extent.height,
                     copy.image_subresource.num_layers * copy.image_extent.depth,
                     output.subspan(output_offset));

            const u32 aligned_plane_dim = Common::AlignUp(copy.image_extent.width, 4) *
                                          Common::AlignUp(copy.image_extent.height, 4);

            copy.buffer_size =
                (aligned_plane_dim * copy.image_extent.depth * copy.image_subresource.num_layers) /
                bpp_div;
            output_offset += static_cast<u32>(copy.buffer_size);
        } else {
            const Extent3D image_extent{
                .width = copy.image_extent.width,
                .height = copy.image_extent.height * copy.image_subresource.num_layers,
                .depth = copy.image_extent.depth,
            };
            DecompressBCn(input_offset, output.subspan(output_offset), image_extent, info.format);
            output_offset += copy.image_extent.width * copy.image_extent.height *
                             copy.image_subresource.num_layers *
                             ConvertedBytesPerBlock(info.format);
        }
    }
}

boost::container::small_vector<BufferImageCopy, 16> FullDownloadCopies(const ImageInfo& info) {
    const Extent3D size = info.size;
    const u32 bytes_per_block = BytesPerBlock(info.format);
    if (info.type == ImageType::Linear) {
        ASSERT(info.pitch % bytes_per_block == 0);
        return {{
            .buffer_offset = 0,
            .buffer_size = static_cast<size_t>(info.pitch) * size.height,
            .buffer_row_length = info.pitch / bytes_per_block,
            .buffer_image_height = size.height,
            .image_subresource =
                {
                    .base_level = 0,
                    .base_layer = 0,
                    .num_layers = 1,
                },
            .image_offset = {0, 0, 0},
            .image_extent = size,
        }};
    }
    UNIMPLEMENTED_IF(info.tile_width_spacing > 0);

    const s32 num_layers = info.resources.layers;
    const s32 num_levels = info.resources.levels;
    const Extent2D tile_size = DefaultBlockSize(info.format);

    u32 host_offset = 0;

    boost::container::small_vector<BufferImageCopy, 16> copies(num_levels);
    for (s32 level = 0; level < num_levels; ++level) {
        const Extent3D level_size = AdjustMipSize(size, level);
        const u32 num_blocks_per_layer = NumBlocks(level_size, tile_size);
        const u32 host_bytes_per_level = num_blocks_per_layer * bytes_per_block * num_layers;
        copies[level] = BufferImageCopy{
            .buffer_offset = host_offset,
            .buffer_size = host_bytes_per_level,
            .buffer_row_length = level_size.width,
            .buffer_image_height = level_size.height,
            .image_subresource =
                {
                    .base_level = level,
                    .base_layer = 0,
                    .num_layers = info.resources.layers,
                },
            .image_offset = {0, 0, 0},
            .image_extent = level_size,
        };
        host_offset += host_bytes_per_level;
    }
    return copies;
}

Extent3D MipSize(Extent3D size, u32 level) {
    return AdjustMipSize(size, level);
}

Extent3D MipBlockSize(const ImageInfo& info, u32 level) {
    const LevelInfo level_info = MakeLevelInfo(info);
    const Extent2D tile_size = DefaultBlockSize(info.format);
    const Extent3D level_size = AdjustMipSize(info.size, level);
    const Extent3D num_tiles = AdjustTileSize(level_size, tile_size);
    return AdjustMipBlockSize(num_tiles, level_info.block, level);
}

boost::container::small_vector<SwizzleParameters, 16> FullUploadSwizzles(const ImageInfo& info) {
    const Extent2D tile_size = DefaultBlockSize(info.format);
    if (info.type == ImageType::Linear) {
        return {SwizzleParameters{
            .num_tiles = AdjustTileSize(info.size, tile_size),
            .block = {},
            .buffer_offset = 0,
            .level = 0,
        }};
    }
    const LevelInfo level_info = MakeLevelInfo(info);
    const Extent3D size = info.size;
    const s32 num_levels = info.resources.levels;

    u32 guest_offset = 0;
    boost::container::small_vector<SwizzleParameters, 16> params(num_levels);
    for (s32 level = 0; level < num_levels; ++level) {
        const Extent3D level_size = AdjustMipSize(size, level);
        const Extent3D num_tiles = AdjustTileSize(level_size, tile_size);
        const Extent3D block = AdjustMipBlockSize(num_tiles, level_info.block, level);
        params[level] = SwizzleParameters{
            .num_tiles = num_tiles,
            .block = block,
            .buffer_offset = guest_offset,
            .level = level,
        };
        guest_offset += CalculateLevelSize(level_info, level);
    }
    return params;
}

void SwizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const ImageInfo& info,
                  std::span<const BufferImageCopy> copies, std::span<const u8> memory,
                  Common::ScratchBuffer<u8>& tmp_buffer) {
    const bool is_pitch_linear = info.type == ImageType::Linear;
    for (const BufferImageCopy& copy : copies) {
        if (is_pitch_linear) {
            SwizzlePitchLinearImage(gpu_memory, gpu_addr, info, copy, memory);
        } else {
            SwizzleBlockLinearImage(gpu_memory, gpu_addr, info, copy, memory, tmp_buffer);
        }
    }
}

bool IsBlockLinearSizeCompatible(const ImageInfo& lhs, const ImageInfo& rhs, u32 lhs_level,
                                 u32 rhs_level, bool strict_size) noexcept {
    ASSERT(lhs.type != ImageType::Linear);
    ASSERT(rhs.type != ImageType::Linear);
    if (strict_size) {
        const Extent3D lhs_size = AdjustMipSize(lhs.size, lhs_level);
        const Extent3D rhs_size = AdjustMipSize(rhs.size, rhs_level);
        return lhs_size.width == rhs_size.width && lhs_size.height == rhs_size.height;
    } else {
        const Extent3D lhs_size = BlockLinearAlignedSize(lhs, lhs_level);
        const Extent3D rhs_size = BlockLinearAlignedSize(rhs, rhs_level);
        return lhs_size.width == rhs_size.width && lhs_size.height == rhs_size.height;
    }
}

bool IsBlockLinearSizeCompatibleBPPRelaxed(const ImageInfo& lhs, const ImageInfo& rhs,
                                           u32 lhs_level, u32 rhs_level) noexcept {
    ASSERT(lhs.type != ImageType::Linear);
    ASSERT(rhs.type != ImageType::Linear);
    const auto lhs_bpp = BytesPerBlock(lhs.format);
    const auto rhs_bpp = BytesPerBlock(rhs.format);
    const Extent3D lhs_size = AdjustMipSize(lhs.size, lhs_level);
    const Extent3D rhs_size = AdjustMipSize(rhs.size, rhs_level);
    return Common::AlignUpLog2(lhs_size.width * lhs_bpp, GOB_SIZE_X_SHIFT) ==
               Common::AlignUpLog2(rhs_size.width * rhs_bpp, GOB_SIZE_X_SHIFT) &&
           Common::AlignUpLog2(lhs_size.height, GOB_SIZE_Y_SHIFT) ==
               Common::AlignUpLog2(rhs_size.height, GOB_SIZE_Y_SHIFT);
}

bool IsPitchLinearSameSize(const ImageInfo& lhs, const ImageInfo& rhs, bool strict_size) noexcept {
    ASSERT(lhs.type == ImageType::Linear);
    ASSERT(rhs.type == ImageType::Linear);
    if (strict_size) {
        return lhs.size.width == rhs.size.width && lhs.size.height == rhs.size.height;
    } else {
        const Extent2D lhs_size = PitchLinearAlignedSize(lhs);
        const Extent2D rhs_size = PitchLinearAlignedSize(rhs);
        return lhs_size == rhs_size;
    }
}

std::optional<OverlapResult> ResolveOverlap(const ImageInfo& new_info, GPUVAddr gpu_addr,
                                            VAddr cpu_addr, const ImageBase& overlap,
                                            bool strict_size, bool broken_views, bool native_bgr) {
    ASSERT(new_info.type != ImageType::Linear);
    ASSERT(overlap.info.type != ImageType::Linear);
    if (!IsLayerStrideCompatible(new_info, overlap.info)) {
        return std::nullopt;
    }
    if (!IsViewCompatible(overlap.info.format, new_info.format, broken_views, native_bgr)) {
        return std::nullopt;
    }
    if (gpu_addr == overlap.gpu_addr) {
        const std::optional solution = ResolveOverlapEqualAddress(new_info, overlap, strict_size);
        if (!solution) {
            return std::nullopt;
        }
        return OverlapResult{
            .gpu_addr = gpu_addr,
            .cpu_addr = cpu_addr,
            .resources = *solution,
        };
    }
    if (overlap.gpu_addr > gpu_addr) {
        return ResolveOverlapRightAddress(new_info, gpu_addr, cpu_addr, overlap, strict_size);
    }
    // if overlap.gpu_addr < gpu_addr
    return ResolveOverlapLeftAddress(new_info, gpu_addr, cpu_addr, overlap, strict_size);
}

bool IsLayerStrideCompatible(const ImageInfo& lhs, const ImageInfo& rhs) {
    // If either of the layer strides is zero, we can assume they are compatible
    // These images generally come from rendertargets
    if (lhs.layer_stride == 0) {
        return true;
    }
    if (rhs.layer_stride == 0) {
        return true;
    }
    // It's definitely compatible if the layer stride matches
    if (lhs.layer_stride == rhs.layer_stride) {
        return true;
    }
    // Although we also have to compare for cases where it can be unaligned
    // This can happen if the image doesn't have layers, so the stride is not aligned
    if (lhs.maybe_unaligned_layer_stride == rhs.maybe_unaligned_layer_stride) {
        return true;
    }
    return false;
}

std::optional<SubresourceBase> FindSubresource(const ImageInfo& candidate, const ImageBase& image,
                                               GPUVAddr candidate_addr, RelaxedOptions options,
                                               bool broken_views, bool native_bgr) {
    const std::optional<SubresourceBase> base = image.TryFindBase(candidate_addr);
    if (!base) {
        return std::nullopt;
    }
    const ImageInfo& existing = image.info;
    if (True(options & RelaxedOptions::Format)) {
        // Format checking is relaxed, but we still have to check for matching bytes per block.
        // This avoids creating a view for blits on UE4 titles where formats with different bytes
        // per block are aliased.
        if (BytesPerBlock(existing.format) != BytesPerBlock(candidate.format)) {
            return std::nullopt;
        }
    } else {
        // Format compatibility is not relaxed, ensure we are creating a view on a compatible format
        if (!IsViewCompatible(existing.format, candidate.format, broken_views, native_bgr)) {
            return std::nullopt;
        }
    }
    if (!IsLayerStrideCompatible(existing, candidate)) {
        return std::nullopt;
    }
    if (existing.type != candidate.type) {
        return std::nullopt;
    }
    if (False(options & RelaxedOptions::Samples) && existing.num_samples != candidate.num_samples) {
        return std::nullopt;
    }
    if (existing.resources.levels < candidate.resources.levels + base->level) {
        return std::nullopt;
    }
    if (existing.type == ImageType::e3D) {
        const u32 mip_depth = std::max(1U, existing.size.depth << base->level);
        if (mip_depth < candidate.size.depth + base->layer) {
            return std::nullopt;
        }
    } else if (existing.resources.layers < candidate.resources.layers + base->layer) {
        return std::nullopt;
    }
    const bool strict_size = False(options & RelaxedOptions::Size);
    if (!IsBlockLinearSizeCompatible(existing, candidate, base->level, 0, strict_size)) {
        return std::nullopt;
    }
    // TODO: compare block sizes
    return base;
}

bool IsSubresource(const ImageInfo& candidate, const ImageBase& image, GPUVAddr candidate_addr,
                   RelaxedOptions options, bool broken_views, bool native_bgr) {
    return FindSubresource(candidate, image, candidate_addr, options, broken_views, native_bgr)
        .has_value();
}

bool IsSubCopy(const ImageInfo& candidate, const ImageBase& image, GPUVAddr candidate_addr) {
    const std::optional<SubresourceBase> base = image.TryFindBase(candidate_addr);
    if (!base) {
        return false;
    }
    const ImageInfo& existing = image.info;
    if (existing.resources.levels < candidate.resources.levels + base->level) {
        return false;
    }
    if (existing.type == ImageType::e3D) {
        const u32 mip_depth = std::max(1U, existing.size.depth << base->level);
        if (mip_depth < candidate.size.depth + base->layer) {
            return false;
        }
    } else {
        if (existing.resources.layers < candidate.resources.layers + base->layer) {
            return false;
        }
    }
    if (!IsBlockLinearSizeCompatibleBPPRelaxed(existing, candidate, base->level, 0)) {
        return false;
    }
    return true;
}

void DeduceBlitImages(ImageInfo& dst_info, ImageInfo& src_info, const ImageBase* dst,
                      const ImageBase* src) {
    const auto original_dst_format = dst_info.format;
    if (src && GetFormatType(src->info.format) != SurfaceType::ColorTexture) {
        src_info.format = src->info.format;
    }
    if (dst && GetFormatType(dst->info.format) != SurfaceType::ColorTexture) {
        dst_info.format = dst->info.format;
    }
    if (src && GetFormatType(src->info.format) != SurfaceType::ColorTexture) {
        dst_info.format = src->info.format;
    }
    if (dst && GetFormatType(dst->info.format) != SurfaceType::ColorTexture) {
        if (src) {
            if (GetFormatType(src->info.format) == SurfaceType::ColorTexture) {
                dst_info.format = original_dst_format;
            }
        } else {
            src_info.format = dst->info.format;
        }
    }
}

u32 MapSizeBytes(const ImageBase& image) {
    if (True(image.flags & ImageFlagBits::AcceleratedUpload)) {
        return image.guest_size_bytes;
    } else if (True(image.flags & ImageFlagBits::Converted)) {
        return image.converted_size_bytes;
    } else {
        return image.unswizzled_size_bytes;
    }
}

static_assert(CalculateLevelSize(LevelInfo{{1920, 1080, 1}, {0, 2, 0}, {1, 1}, 2, 0}, 0) ==
              0x7f8000);
static_assert(CalculateLevelSize(LevelInfo{{32, 32, 1}, {0, 0, 4}, {1, 1}, 4, 0}, 0) == 0x40000);

static_assert(CalculateLevelSize(LevelInfo{{128, 8, 1}, {0, 4, 0}, {1, 1}, 4, 0}, 0) == 0x40000);

static_assert(CalculateLevelOffset(PixelFormat::R8_SINT, {1920, 1080, 1}, {0, 2, 0}, 0, 7) ==
              0x2afc00);
static_assert(CalculateLevelOffset(PixelFormat::ASTC_2D_12X12_UNORM, {8192, 4096, 1}, {0, 2, 0}, 0,
                                   12) == 0x50d200);

static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 0) ==
              0);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 1) ==
              0x400000);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 2) ==
              0x500000);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 3) ==
              0x540000);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 4) ==
              0x550000);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 5) ==
              0x554000);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 6) ==
              0x555000);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 7) ==
              0x555400);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 8) ==
              0x555600);
static_assert(CalculateLevelOffset(PixelFormat::A8B8G8R8_UNORM, {1024, 1024, 1}, {0, 4, 0}, 0, 9) ==
              0x555800);

constexpr u32 ValidateLayerSize(PixelFormat format, u32 width, u32 height, u32 block_height,
                                u32 tile_width_spacing, u32 level) {
    const Extent3D size{width, height, 1};
    const Extent3D block{0, block_height, 0};
    const u32 offset = CalculateLevelOffset(format, size, block, tile_width_spacing, level);
    return AlignLayerSize(offset, size, block, DefaultBlockHeight(format), tile_width_spacing);
}

static_assert(ValidateLayerSize(PixelFormat::ASTC_2D_12X12_UNORM, 8192, 4096, 2, 0, 12) ==
              0x50d800);
static_assert(ValidateLayerSize(PixelFormat::A8B8G8R8_UNORM, 1024, 1024, 2, 0, 10) == 0x556000);
static_assert(ValidateLayerSize(PixelFormat::BC3_UNORM, 128, 128, 2, 0, 8) == 0x6000);

static_assert(ValidateLayerSize(PixelFormat::A8B8G8R8_UNORM, 518, 572, 4, 3, 1) == 0x190000,
              "Tile width spacing is not working");
static_assert(ValidateLayerSize(PixelFormat::BC5_UNORM, 1024, 1024, 3, 4, 11) == 0x160000,
              "Compressed tile width spacing is not working");

} // namespace VideoCommon