1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739
// Copyright (c) 2016 The vulkano developers
// Licensed under the Apache License, Version 2.0
// <LICENSE-APACHE or
// https://www.apache.org/licenses/LICENSE-2.0> or the MIT
// license <LICENSE-MIT or https://opensource.org/licenses/MIT>,
// at your option. All files in the project carrying such
// notice may not be copied, modified, or distributed except
// according to those terms.
//! In Vulkan, suballocation of [`DeviceMemory`] is left to the application, because every
//! application has slightly different needs and one can not incorporate an allocator into the
//! driver that would perform well in all cases. Vulkano stays true to this sentiment, but aims to
//! reduce the burden on the user as much as possible. You have a toolbox of [suballocators] to
//! choose from that cover all allocation algorithms, which you can compose into any kind of
//! [hierarchy] you wish. This way you have maximum flexibility while still only using a few
//! `DeviceMemory` blocks and not writing any of the very error-prone code.
//!
//! If you just want to allocate memory and don't have any special needs, look no further than the
//! [`StandardMemoryAllocator`].
//!
//! # Why not just allocate `DeviceMemory`?
//!
//! But the driver has an allocator! Otherwise you wouldn't be able to allocate `DeviceMemory`,
//! right? Indeed, but that allocation is very expensive. Not only that, there is also a pretty low
//! limit on the number of allocations by the drivers. See, everything in Vulkan tries to keep you
//! away from allocating `DeviceMemory` too much. These limits are used by the implementation to
//! optimize on its end, while the application optimizes on the other end.
//!
//! # Alignment
//!
//! At the end of the day, memory needs to be backed by hardware somehow. A *memory cell* stores a
//! single *bit*, bits are grouped into *bytes* and bytes are grouped into *words*. Intuitively, it
//! should make sense that accessing single bits at a time would be very inefficient. That is why
//! computers always access a whole word of memory at once, at least. That means that if you tried
//! to do an unaligned access, you would need to access twice the number of memory locations.
//!
//! Example aligned access, performing bitwise NOT on the (64-bit) word at offset 0x08:
//!
//! ```plain
//! | 08 | 10 | 18
//! ----+-------------------------+-------------------------+----
//! ••• | 35 35 35 35 35 35 35 35 | 01 23 45 67 89 ab cd ef | •••
//! ----+-------------------------+-------------------------+----
//! , | ,
//! +------------|------------+
//! ' v '
//! ----+-------------------------+-------------------------+----
//! ••• | ca ca ca ca ca ca ca ca | 01 23 45 67 89 ab cd ef | •••
//! ----+-------------------------+-------------------------+----
//! ```
//!
//! Same example as above, but this time unaligned with a word at offset 0x0a:
//!
//! ```plain
//! | 08 0a | 10 | 18
//! ----+-------------------------+-------------------------+----
//! ••• | cd ef 35 35 35 35 35 35 | 35 35 01 23 45 67 89 ab | •••
//! ----+-------------------------+-------------------------+----
//! , | ,
//! +------------|------------+
//! ' v '
//! ----+-------------------------+-------------------------+----
//! ••• | cd ef ca ca ca ca ca ca | ca ca 01 23 45 67 89 ab | •••
//! ----+-------------------------+-------------------------+----
//! ```
//!
//! As you can see, in the unaligned case the hardware would need to read both the word at offset
//! 0x08 and the word at the offset 0x10 and then shift the bits from one register into the other.
//! Safe to say it should to be avoided, and this is why we need alignment. This example also goes
//! to show how inefficient unaligned writes are. Say you pieced together your word as described,
//! and now you want to perform the bitwise NOT and write the result back. Difficult, isn't it?
//! That's due to the fact that even though the chunks occupy different ranges in memory, they are
//! still said to *alias* each other, because if you try to write to one memory location, you would
//! be overwriting 2 or more different chunks of data.
//!
//! ## Pages
//!
//! It doesn't stop at the word, though. Words are further grouped into *pages*. These are
//! typically power-of-two multiples of the word size, much like words are typically powers of two
//! themselves. You can easily extend the concepts from the previous examples to pages if you think
//! of the examples as having a page size of 1 word. Two resources are said to alias if they share
//! a page, and therefore should be aligned to the page size. What the page size is depends on the
//! context, and a computer might have multiple different ones for different parts of hardware.
//!
//! ## Memory requirements
//!
//! A Vulkan device might have any number of reasons it would want certain alignments for certain
//! resources. For example, the device might have different caches for different types of
//! resources, which have different page sizes. Maybe the device wants to store images in some
//! other cache compared to buffers which needs different alignment. Or maybe images of different
//! layouts require different alignment, or buffers with different usage/mapping do. The specifics
//! don't matter in the end, this just goes to illustrate the point. This is why memory
//! requirements in Vulkan vary not only with the Vulkan implementation, but also with the type of
//! resource.
//!
//! ## Buffer-image granularity
//!
//! This unfortunately named granularity is the page size which a linear resource neighboring a
//! non-linear resource must be aligned to in order for them not to alias. The difference between
//! the memory requirements of the individual resources and the [buffer-image granularity] is that
//! the memory requirements only apply to the resource they are for, while the buffer-image
//! granularity applies to two neighboring resources. For example, you might create two buffers,
//! which might have two different memory requirements, but as long as those are satisfied, you can
//! put these buffers cheek to cheek. On the other hand, if one of them is an (optimal layout)
//! image, then they must not share any page, whose size is given by this granularity. The Vulkan
//! implementation can use this for additional optimizations if it needs to, or report a
//! granularity of 1.
//!
//! # Fragmentation
//!
//! Memory fragmentation refers to the wastage of memory that results from alignment requirements
//! and/or dynamic memory allocation. As such, some level of fragmentation is always going to be
//! inevitable. Different allocation algorithms each have their own characteristics and trade-offs
//! in relation to fragmentation.
//!
//! ## Internal Fragmentation
//!
//! This type of fragmentation arises from alignment requirements. These might be imposed by the
//! Vulkan implementation or the application itself.
//!
//! Say for example your allocations need to be aligned to 64B, then any allocation whose size is
//! not a multiple of the alignment will need padding at the end:
//!
//! ```plain
//! | 0x040 | 0x080 | 0x0c0 | 0x100
//! ----+------------------+------------------+------------------+--------
//! | ############ | ################ | ######## | #######
//! ••• | ### 48 B ### | ##### 64 B ##### | # 32 B # | ### •••
//! | ############ | ################ | ######## | #######
//! ----+------------------+------------------+------------------+--------
//! ```
//!
//! If this alignment is imposed by the Vulkan implementation, then there's nothing one can do
//! about this. Simply put, that space is unusable. One also shouldn't want to do anything about
//! it, since these requirements have very good reasons, as described in further detail in previous
//! sections. They prevent resources from aliasing so that performance is optimal.
//!
//! It might seem strange that the application would want to cause internal fragmentation itself,
//! but this is often a good trade-off to reduce or even completely eliminate external
//! fragmentation. Internal fragmentation is very predictable, which makes it easier to deal with.
//!
//! ## External fragmentation
//!
//! With external fragmentation, what happens is that while the allocations might be using their
//! own memory totally efficiently, the way they are arranged in relation to each other would
//! prevent a new contiguous chunk of memory to be allocated even though there is enough free space
//! left. That is why this fragmentation is said to be external to the allocations. Also, the
//! allocations together with the fragments in-between add overhead both in terms of space and time
//! to the allocator, because it needs to keep track of more things overall.
//!
//! As an example, take these 4 allocations within some block, with the rest of the block assumed
//! to be full:
//!
//! ```plain
//! +-----+-------------------+-------+-----------+-- - - --+
//! | | | | | |
//! | A | B | C | D | ••• |
//! | | | | | |
//! +-----+-------------------+-------+-----------+-- - - --+
//! ```
//!
//! The allocations were all done in order, and naturally there is no fragmentation at this point.
//! Now if we free B and D, since these are done out of order, we will be left with holes between
//! the other allocations, and we won't be able to fit allocation E anywhere:
//!
//! ```plain
//! +-----+-------------------+-------+-----------+-- - - --+ +-------------------------+
//! | | | | | | ? | |
//! | A | | C | | ••• | <== | E |
//! | | | | | | | |
//! +-----+-------------------+-------+-----------+-- - - --+ +-------------------------+
//! ```
//!
//! So fine, we use a different block for E, and just use this block for allocations that fit:
//!
//! ```plain
//! +-----+---+-----+---------+-------+-----+-----+-- - - --+
//! | | | | | | | | |
//! | A | H | I | J | C | F | G | ••• |
//! | | | | | | | | |
//! +-----+---+-----+---------+-------+-----+-----+-- - - --+
//! ```
//!
//! Sure, now let's free some shall we? And voilà, the problem just became much worse:
//!
//! ```plain
//! +-----+---+-----+---------+-------+-----+-----+-- - - --+
//! | | | | | | | | |
//! | A | | I | J | | F | | ••• |
//! | | | | | | | | |
//! +-----+---+-----+---------+-------+-----+-----+-- - - --+
//! ```
//!
//! # Leakage
//!
//! Memory leaks happen when allocations are kept alive past their shelf life. This most often
//! occurs because of [cyclic references]. If you have structures that have cycles, then make sure
//! you read the documentation for [`Arc`]/[`Rc`] carefully to avoid memory leaks. You can also
//! introduce memory leaks willingly by using [`mem::forget`] or [`Box::leak`] to name a few. In
//! all of these examples the memory can never be reclaimed, but that doesn't have to be the case
//! for something to be considered a leak. Say for example you have a [region] which you
//! suballocate, and at some point you drop all the suballocations. When that happens, the region
//! can be returned (freed) to the next level up the hierarchy, or it can be reused by another
//! suballocator. But if you happen to keep alive just one suballocation for the duration of the
//! program for instance, then the whole region is also kept as it is for that time (and keep in
//! mind this bubbles up the hierarchy). Therefore, for the program, that memory might be a leak
//! depending on the allocator, because some allocators wouldn't be able to reuse the entire rest
//! of the region. You must always consider the lifetime of your resources when choosing the
//! appropriate allocator.
//!
//! [suballocators]: Suballocator
//! [hierarchy]: Suballocator#memory-hierarchies
//! [buffer-image granularity]: crate::device::Properties::buffer_image_granularity
//! [cyclic references]: Arc#breaking-cycles-with-weak
//! [`Rc`]: std::rc::Rc
//! [`mem::forget`]: std::mem::forget
//! [region]: Suballocator#regions
mod layout;
pub mod suballocator;
use self::{array_vec::ArrayVec, suballocator::Region};
pub use self::{
layout::DeviceLayout,
suballocator::{
AllocationType, BuddyAllocator, BumpAllocator, FreeListAllocator, Suballocation,
Suballocator, SuballocatorError,
},
};
use super::{
DedicatedAllocation, DeviceAlignment, DeviceMemory, ExternalMemoryHandleTypes,
MemoryAllocateFlags, MemoryAllocateInfo, MemoryMapInfo, MemoryProperties, MemoryPropertyFlags,
MemoryRequirements, MemoryType,
};
use crate::{
device::{Device, DeviceOwned},
instance::InstanceOwnedDebugWrapper,
DeviceSize, Validated, Version, VulkanError,
};
use ash::vk::MAX_MEMORY_TYPES;
use parking_lot::Mutex;
use std::{
error::Error,
fmt::{Debug, Display, Error as FmtError, Formatter},
mem,
ops::BitOr,
ptr,
sync::Arc,
};
/// General-purpose memory allocators which allocate from any memory type dynamically as needed.
///
/// # Safety
///
/// - `allocate`, `allocate_from_type` and `allocate_dedicated` must return a memory block that is
/// in bounds of its device memory.
/// - `allocate` and `allocate_from_type` must return a memory block that doesn't alias any other
/// currently allocated memory blocks:
/// - Two currently allocated memory blocks must not share any memory locations, meaning that the
/// intersection of the byte ranges of the two memory blocks must be empty.
/// - Two neighboring currently allocated memory blocks must not share any [page] whose size is
/// given by the [buffer-image granularity], unless either both were allocated with
/// [`AllocationType::Linear`] or both were allocated with [`AllocationType::NonLinear`].
/// - For all [host-visible] memory types that are not [host-coherent], all memory blocks must be
/// aligned to the [non-coherent atom size].
/// - The size does **not** have to be padded to the alignment. That is, as long the offset is
/// aligned and the memory blocks don't share any memory locations, a memory block is not
/// considered to alias another even if the padded size shares memory locations with another
/// memory block.
/// - A memory block must stay allocated until either `deallocate` is called on it or the allocator
/// is dropped. If the allocator is cloned, it must produce the same allocator, and memory blocks
/// must stay allocated until either `deallocate` is called on the memory block using any of the
/// clones or all of the clones have been dropped.
///
/// [page]: self#pages
/// [buffer-image granularity]: self#buffer-image-granularity
/// [host-visible]: MemoryPropertyFlags::HOST_VISIBLE
/// [host-coherent]: MemoryPropertyFlags::HOST_COHERENT
/// [non-coherent atom size]: crate::device::Properties::non_coherent_atom_size
pub unsafe trait MemoryAllocator: DeviceOwned + Send + Sync + 'static {
/// Finds the most suitable memory type index in `memory_type_bits` using the given `filter`.
/// Returns [`None`] if the requirements are too strict and no memory type is able to satisfy
/// them.
fn find_memory_type_index(
&self,
memory_type_bits: u32,
filter: MemoryTypeFilter,
) -> Option<u32>;
/// Allocates memory from a specific memory type.
///
/// # Arguments
///
/// - `memory_type_index` - The index of the memory type to allocate from.
///
/// - `layout` - The layout of the allocation.
///
/// - `allocation_type` - The type of resources that can be bound to the allocation.
///
/// - `never_allocate` - If `true` then the allocator should never allocate `DeviceMemory`,
/// instead only suballocate from existing blocks.
fn allocate_from_type(
&self,
memory_type_index: u32,
layout: DeviceLayout,
allocation_type: AllocationType,
never_allocate: bool,
) -> Result<MemoryAlloc, MemoryAllocatorError>;
/// Allocates memory according to requirements.
///
/// # Arguments
///
/// - `requirements` - Requirements of the resource you want to allocate memory for.
///
/// If you plan to bind this memory directly to a non-sparse resource, then this must
/// correspond to the value returned by either [`RawBuffer::memory_requirements`] or
/// [`RawImage::memory_requirements`] for the respective buffer or image.
///
/// - `allocation_type` - What type of resource this allocation will be used for.
///
/// This should be [`Linear`] for buffers and linear images, and [`NonLinear`] for optimal
/// images. You can not bind memory allocated with the [`Linear`] type to optimal images or
/// bind memory allocated with the [`NonLinear`] type to buffers and linear images. You
/// should never use the [`Unknown`] type unless you have to, as that can be less memory
/// efficient.
///
/// - `dedicated_allocation` - Allows a dedicated allocation to be created.
///
/// You should always fill this field in if you are allocating memory for a non-sparse
/// resource, otherwise the allocator won't be able to create a dedicated allocation if one
/// is required or recommended.
///
/// This argument is silently ignored (treated as `None`) if the device API version is below
/// 1.1 and the [`khr_dedicated_allocation`] extension is not enabled on the device.
///
/// [`RawBuffer::memory_requirements`]: crate::buffer::sys::RawBuffer::memory_requirements
/// [`RawImage::memory_requirements`]: crate::image::sys::RawImage::memory_requirements
/// [`Linear`]: AllocationType::Linear
/// [`NonLinear`]: AllocationType::NonLinear
/// [`Unknown`]: AllocationType::Unknown
/// [`khr_dedicated_allocation`]: crate::device::DeviceExtensions::khr_dedicated_allocation
fn allocate(
&self,
requirements: MemoryRequirements,
allocation_type: AllocationType,
create_info: AllocationCreateInfo,
dedicated_allocation: Option<DedicatedAllocation<'_>>,
) -> Result<MemoryAlloc, MemoryAllocatorError>;
/// Creates an allocation with a whole device memory block dedicated to it.
fn allocate_dedicated(
&self,
memory_type_index: u32,
allocation_size: DeviceSize,
dedicated_allocation: Option<DedicatedAllocation<'_>>,
export_handle_types: ExternalMemoryHandleTypes,
) -> Result<MemoryAlloc, MemoryAllocatorError>;
/// Deallocates the given `allocation`.
///
/// # Safety
///
/// - `allocation` must refer to a **currently allocated** allocation of `self`.
unsafe fn deallocate(&self, allocation: MemoryAlloc);
}
impl Debug for dyn MemoryAllocator {
fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), FmtError> {
f.debug_struct("MemoryAllocator").finish_non_exhaustive()
}
}
/// Describes what memory property flags are required, preferred and not preferred when picking a
/// memory type index.
///
/// # Examples
///
/// For resources that the device frequently accesses, e.g. textures, render targets, or
/// intermediary buffers, you want device-local memory without any host access:
///
/// ```
/// # use vulkano::{
/// # image::{Image, ImageCreateInfo, ImageUsage},
/// # memory::allocator::{AllocationCreateInfo, MemoryTypeFilter},
/// # };
/// #
/// # let memory_allocator: std::sync::Arc<vulkano::memory::allocator::StandardMemoryAllocator> = return;
/// # let format = return;
/// # let extent = return;
/// #
/// let texture = Image::new(
/// memory_allocator.clone(),
/// ImageCreateInfo {
/// format,
/// extent,
/// usage: ImageUsage::TRANSFER_DST | ImageUsage::SAMPLED,
/// ..Default::default()
/// },
/// AllocationCreateInfo {
/// memory_type_filter: MemoryTypeFilter::PREFER_DEVICE,
/// ..Default::default()
/// },
/// )
/// .unwrap();
/// ```
///
/// For staging, the resource is only ever written to sequentially. Also, since the device will
/// only read the staging resourse once, it would yield no benefit to place it in device-local
/// memory, in fact it would be wasteful. Therefore, it's best to put it in host-local memory:
///
/// ```
/// # use vulkano::{
/// # buffer::{Buffer, BufferCreateInfo, BufferUsage},
/// # memory::allocator::{AllocationCreateInfo, MemoryTypeFilter},
/// # };
/// #
/// # let memory_allocator: std::sync::Arc<vulkano::memory::allocator::StandardMemoryAllocator> = return;
/// #
/// let staging_buffer = Buffer::new_sized(
/// memory_allocator.clone(),
/// BufferCreateInfo {
/// usage: BufferUsage::TRANSFER_SRC,
/// ..Default::default()
/// },
/// AllocationCreateInfo {
/// memory_type_filter: MemoryTypeFilter::PREFER_HOST |
/// MemoryTypeFilter::HOST_SEQUENTIAL_WRITE,
/// ..Default::default()
/// },
/// )
/// .unwrap();
/// #
/// # let staging_buffer: vulkano::buffer::Subbuffer<u32> = staging_buffer;
/// ```
///
/// For resources that the device accesses directly, aka a buffer/image used for anything other
/// than transfers, it's probably best to put it in device-local memory:
///
/// ```
/// # use vulkano::{
/// # buffer::{Buffer, BufferCreateInfo, BufferUsage},
/// # memory::allocator::{AllocationCreateInfo, MemoryTypeFilter},
/// # };
/// #
/// # let memory_allocator: std::sync::Arc<vulkano::memory::allocator::StandardMemoryAllocator> = return;
/// #
/// let uniform_buffer = Buffer::new_sized(
/// memory_allocator.clone(),
/// BufferCreateInfo {
/// usage: BufferUsage::UNIFORM_BUFFER,
/// ..Default::default()
/// },
/// AllocationCreateInfo {
/// memory_type_filter: MemoryTypeFilter::PREFER_DEVICE |
/// MemoryTypeFilter::HOST_SEQUENTIAL_WRITE,
/// ..Default::default()
/// },
/// )
/// .unwrap();
/// #
/// # let uniform_buffer: vulkano::buffer::Subbuffer<u32> = uniform_buffer;
/// ```
///
/// For readback, e.g. getting the results of a compute shader back to the host:
///
/// ```
/// # use vulkano::{
/// # buffer::{Buffer, BufferCreateInfo, BufferUsage},
/// # memory::allocator::{AllocationCreateInfo, MemoryTypeFilter},
/// # };
/// #
/// # let memory_allocator: std::sync::Arc<vulkano::memory::allocator::StandardMemoryAllocator> = return;
/// #
/// let readback_buffer = Buffer::new_sized(
/// memory_allocator.clone(),
/// BufferCreateInfo {
/// usage: BufferUsage::TRANSFER_DST,
/// ..Default::default()
/// },
/// AllocationCreateInfo {
/// memory_type_filter: MemoryTypeFilter::PREFER_HOST |
/// MemoryTypeFilter::HOST_RANDOM_ACCESS,
/// ..Default::default()
/// },
/// )
/// .unwrap();
/// #
/// # let readback_buffer: vulkano::buffer::Subbuffer<u32> = readback_buffer;
/// ```
#[derive(Clone, Copy, Debug, Default, PartialEq, Eq)]
pub struct MemoryTypeFilter {
pub required_flags: MemoryPropertyFlags,
pub preferred_flags: MemoryPropertyFlags,
pub not_preferred_flags: MemoryPropertyFlags,
}
impl MemoryTypeFilter {
/// Prefers picking a memory type with the [`DEVICE_LOCAL`] flag.
///
/// Memory being device-local means that it is fastest to access for the device. However,
/// for dedicated GPUs, getting memory in and out of VRAM requires to go through the PCIe bus,
/// which is very slow and should therefore only be done when necessary.
///
/// This filter is best suited for anything that the host doesn't access, but the device
/// accesses frequently. For example textures, render targets, and intermediary buffers.
///
/// For memory that the host does access but less frequently than the device, such as updating
/// a uniform buffer each frame, device-local memory may also be preferred. In this case,
/// because the memory is only written once before being consumed by the device and becoming
/// outdated, it doesn't matter that the data is potentially transferred over the PCIe bus
/// since it only happens once. Since this is only a preference, if you have some requirements
/// such as the memory being [`HOST_VISIBLE`], those requirements will take precendence.
///
/// For memory that the host doesn't access, and the device doesn't access directly, you may
/// still prefer device-local memory if the memory is used regularly. For instance, an image
/// that is swapped each frame.
///
/// Keep in mind that for implementations with unified memory, there's only system RAM. That
/// means that even if the implementation reports a memory type that is not `HOST_VISIBLE` and
/// is `DEVICE_LOCAL`, its still the same unified memory as all other memory. However, it may
/// still be faster to access. On the other hand, some such implementations don't report any
/// memory types that are not `HOST_VISIBLE`, which makes sense since it's all system RAM. In
/// that case you wouldn't need to do staging.
///
/// Don't use this together with [`PREFER_HOST`], that makes no sense. If you need host access,
/// make sure you combine this with either the [`HOST_SEQUENTIAL_WRITE`] or
/// [`HOST_RANDOM_ACCESS`] filter.
///
/// [`DEVICE_LOCAL`]: MemoryPropertyFlags::DEVICE_LOCAL
/// [`HOST_VISIBLE`]: MemoryPropertyFlags::HOST_VISIBLE
/// [`PREFER_HOST`]: Self::PREFER_HOST
/// [`HOST_SEQUENTIAL_WRITE`]: Self::HOST_SEQUENTIAL_WRITE
/// [`HOST_RANDOM_ACCESS`]: Self::HOST_RANDOM_ACCESS
pub const PREFER_DEVICE: Self = Self {
required_flags: MemoryPropertyFlags::empty(),
preferred_flags: MemoryPropertyFlags::DEVICE_LOCAL,
not_preferred_flags: MemoryPropertyFlags::empty(),
};
/// Prefers picking a memory type without the [`DEVICE_LOCAL`] flag.
///
/// This option is best suited for resources that the host does access, but device doesn't
/// access directly, such as staging buffers and readback buffers.
///
/// For memory that the host does access but less frequently than the device, such as updating
/// a uniform buffer each frame, you may still get away with using host-local memory if the
/// updates are small and local enough. In such cases, the memory should be able to be quickly
/// cached by the device, such that the data potentially being transferred over the PCIe bus
/// wouldn't matter.
///
/// For memory that the host doesn't access, and the device doesn't access directly, you may
/// still prefer host-local memory if the memory is rarely used, such as for manually paging
/// parts of device-local memory out in order to free up space on the device.
///
/// Don't use this together with [`PREFER_DEVICE`], that makes no sense. If you need host
/// access, make sure you combine this with either the [`HOST_SEQUENTIAL_WRITE`] or
/// [`HOST_RANDOM_ACCESS`] filter.
///
/// [`DEVICE_LOCAL`]: MemoryPropertyFlags::DEVICE_LOCAL
/// [`PREFER_DEVICE`]: Self::PREFER_DEVICE
/// [`HOST_SEQUENTIAL_WRITE`]: Self::HOST_SEQUENTIAL_WRITE
/// [`HOST_RANDOM_ACCESS`]: Self::HOST_RANDOM_ACCESS
pub const PREFER_HOST: Self = Self {
required_flags: MemoryPropertyFlags::empty(),
preferred_flags: MemoryPropertyFlags::empty(),
not_preferred_flags: MemoryPropertyFlags::DEVICE_LOCAL,
};
/// This guarantees picking a memory type that has the [`HOST_VISIBLE`] flag. Using this filter
/// allows the allocator to pick a memory type that is uncached and write-combined, which is
/// ideal for sequential writes. However, this optimization might lead to poor performance for
/// anything else. What counts as a sequential write is any kind of loop that writes memory
/// locations in order, such as iterating over a slice while writing each element in order, or
/// equivalently using [`slice::copy_from_slice`]. Copying sized data also counts, as rustc
/// should write the memory locations in order. If you have a struct, make sure you write it
/// member-by-member.
///
/// Example use cases include staging buffers, as well as any other kind of buffer that you
/// only write to from the host, like a uniform or vertex buffer.
///
/// Don't use this together with [`HOST_RANDOM_ACCESS`], that makes no sense. If you do both a
/// sequential write and read or random access, then you should use `HOST_RANDOM_ACCESS`
/// instead. However, you could also consider using different allocations for the two purposes
/// to get the most performance out, if that's possible.
///
/// [`HOST_VISIBLE`]: MemoryPropertyFlags::HOST_VISIBLE
/// [`HOST_COHERENT`]: MemoryPropertyFlags::HOST_COHERENT
/// [`HOST_RANDOM_ACCESS`]: Self::HOST_RANDOM_ACCESS
pub const HOST_SEQUENTIAL_WRITE: Self = Self {
required_flags: MemoryPropertyFlags::HOST_VISIBLE,
preferred_flags: MemoryPropertyFlags::empty(),
not_preferred_flags: MemoryPropertyFlags::HOST_CACHED,
};
/// This guarantees picking a memory type that has the [`HOST_VISIBLE`] and [`HOST_CACHED`]
/// flags, which is best suited for readback and/or random access.
///
/// Example use cases include using the device for things other than rendering and getting the
/// results back to the host. That might be compute shading, or image or video manipulation, or
/// screenshotting.
///
/// Don't use this together with [`HOST_SEQUENTIAL_WRITE`], that makes no sense. If you are
/// sure you only ever need to sequentially write to the allocation, then using
/// `HOST_SEQUENTIAL_WRITE` instead will yield better performance.
///
/// [`HOST_VISIBLE`]: MemoryPropertyFlags::HOST_VISIBLE
/// [`HOST_CACHED`]: MemoryPropertyFlags::HOST_CACHED
/// [`HOST_SEQUENTIAL_WRITE`]: Self::HOST_SEQUENTIAL_WRITE
pub const HOST_RANDOM_ACCESS: Self = Self {
required_flags: MemoryPropertyFlags::HOST_VISIBLE.union(MemoryPropertyFlags::HOST_CACHED),
preferred_flags: MemoryPropertyFlags::empty(),
not_preferred_flags: MemoryPropertyFlags::empty(),
};
/// Returns a `MemoryTypeFilter` with none of the flags set.
#[inline]
pub const fn empty() -> Self {
Self {
required_flags: MemoryPropertyFlags::empty(),
preferred_flags: MemoryPropertyFlags::empty(),
not_preferred_flags: MemoryPropertyFlags::empty(),
}
}
/// Returns the union of `self` and `other`.
#[inline]
pub const fn union(self, other: Self) -> Self {
Self {
required_flags: self.required_flags.union(other.required_flags),
preferred_flags: self.preferred_flags.union(other.preferred_flags),
not_preferred_flags: self.not_preferred_flags.union(other.not_preferred_flags),
}
}
}
impl BitOr for MemoryTypeFilter {
type Output = Self;
#[inline]
fn bitor(self, rhs: Self) -> Self::Output {
self.union(rhs)
}
}
/// Parameters to create a new [allocation] using a [memory allocator].
///
/// [allocation]: MemoryAlloc
/// [memory allocator]: MemoryAllocator
#[derive(Clone, Debug)]
pub struct AllocationCreateInfo {
/// Filter used to narrow down the memory type to be selected.
///
/// The default value is [`MemoryTypeFilter::PREFER_DEVICE`].
pub memory_type_filter: MemoryTypeFilter,
/// Allows you to further constrain the possible choices of memory types, by only allowing the
/// memory type indices that have a corresponding bit at the same index set to 1.
///
/// The default value is [`u32::MAX`].
pub memory_type_bits: u32,
/// How eager the allocator should be to allocate [`DeviceMemory`].
///
/// The default value is [`MemoryAllocatePreference::Unknown`].
pub allocate_preference: MemoryAllocatePreference,
pub _ne: crate::NonExhaustive,
}
impl Default for AllocationCreateInfo {
#[inline]
fn default() -> Self {
AllocationCreateInfo {
memory_type_filter: MemoryTypeFilter::PREFER_DEVICE,
memory_type_bits: u32::MAX,
allocate_preference: MemoryAllocatePreference::Unknown,
_ne: crate::NonExhaustive(()),
}
}
}
/// Describes whether allocating [`DeviceMemory`] is desired.
#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
pub enum MemoryAllocatePreference {
/// There is no known preference, let the allocator decide.
Unknown,
/// The allocator should never allocate `DeviceMemory` and should instead only suballocate from
/// existing blocks.
///
/// This option is best suited if you can not afford the overhead of allocating `DeviceMemory`.
NeverAllocate,
/// The allocator should always allocate `DeviceMemory`.
///
/// This option is best suited if you are allocating a long-lived resource that you know could
/// benefit from having a dedicated allocation.
AlwaysAllocate,
}
/// An allocation made using a [memory allocator].
///
/// [memory allocator]: MemoryAllocator
#[derive(Clone, Debug)]
pub struct MemoryAlloc {
/// The underlying block of device memory.
pub device_memory: Arc<DeviceMemory>,
/// The suballocation within the device memory block, or [`None`] if this is a dedicated
/// allocation.
pub suballocation: Option<Suballocation>,
/// An opaque handle identifying the allocation inside the allocator.
pub allocation_handle: AllocationHandle,
}
/// An opaque handle identifying an allocation inside an allocator.
#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
#[cfg_attr(not(doc), repr(transparent))]
pub struct AllocationHandle(*mut ());
unsafe impl Send for AllocationHandle {}
unsafe impl Sync for AllocationHandle {}
impl AllocationHandle {
/// Creates a null `AllocationHandle`.
///
/// Use this if you don't have anything that you need to associate with the allocation.
#[inline]
pub const fn null() -> Self {
AllocationHandle(ptr::null_mut())
}
/// Stores a pointer in an `AllocationHandle`.
///
/// Use this if you want to associate an allocation with some (host) heap allocation.
#[inline]
pub const fn from_ptr(ptr: *mut ()) -> Self {
AllocationHandle(ptr)
}
/// Stores an index inside an `AllocationHandle`.
///
/// Use this if you want to associate an allocation with some index.
#[allow(clippy::useless_transmute)]
#[inline]
pub const fn from_index(index: usize) -> Self {
// SAFETY: `usize` and `*mut ()` have the same layout.
AllocationHandle(unsafe { mem::transmute::<usize, *mut ()>(index) })
}
/// Retrieves a previously-stored pointer from the `AllocationHandle`.
///
/// If this handle hasn't been created using [`from_ptr`] then this will return an invalid
/// pointer, dereferencing which is undefined behavior.
///
/// [`from_ptr`]: Self::from_ptr
#[inline]
pub const fn as_ptr(self) -> *mut () {
self.0
}
/// Retrieves a previously-stored index from the `AllocationHandle`.
///
/// If this handle hasn't been created using [`from_index`] then this will return a bogus
/// result.
///
/// [`from_index`]: Self::from_index
#[allow(clippy::transmutes_expressible_as_ptr_casts)]
#[inline]
pub const fn as_index(self) -> usize {
// SAFETY: `usize` and `*mut ()` have the same layout.
unsafe { mem::transmute::<*mut (), usize>(self.0) }
}
}
/// Error that can be returned when creating an [allocation] using a [memory allocator].
///
/// [allocation]: MemoryAlloc
/// [memory allocator]: MemoryAllocator
#[derive(Clone, Debug)]
pub enum MemoryAllocatorError {
/// Allocating [`DeviceMemory`] failed.
AllocateDeviceMemory(Validated<VulkanError>),
/// Finding a suitable memory type failed.
///
/// This is returned from [`MemoryAllocator::allocate`] when
/// [`MemoryAllocator::find_memory_type_index`] returns [`None`].
FindMemoryType,
/// There is not enough memory in the pool.
///
/// This is returned when using [`MemoryAllocatePreference::NeverAllocate`] and there is not
/// enough memory in the pool.
OutOfPoolMemory,
/// A dedicated allocation is required but was explicitly forbidden.
///
/// This is returned when using [`MemoryAllocatePreference::NeverAllocate`] and the
/// implementation requires a dedicated allocation.
DedicatedAllocationRequired,
/// The block size for the allocator was exceeded.
///
/// This is returned when using [`MemoryAllocatePreference::NeverAllocate`] and the allocation
/// size exceeded the block size for all heaps of suitable memory types.
BlockSizeExceeded,
}
impl Error for MemoryAllocatorError {
fn source(&self) -> Option<&(dyn Error + 'static)> {
match self {
Self::AllocateDeviceMemory(err) => Some(err),
_ => None,
}
}
}
impl Display for MemoryAllocatorError {
fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), FmtError> {
let msg = match self {
Self::AllocateDeviceMemory(_) => "allocating device memory failed",
Self::FindMemoryType => "finding a suitable memory type failed",
Self::OutOfPoolMemory => "the pool doesn't have enough free space",
Self::DedicatedAllocationRequired => {
"a dedicated allocation is required but was explicitly forbidden"
}
Self::BlockSizeExceeded => {
"the allocation size was greater than the block size for all heaps of suitable \
memory types and dedicated allocations were explicitly forbidden"
}
};
f.write_str(msg)
}
}
/// Standard memory allocator intended as a global and general-purpose allocator.
///
/// This type of allocator is what you should always use, unless you know, for a fact, that it is
/// not suited to the task.
///
/// See also [`GenericMemoryAllocator`] for details about the allocation algorithm, and
/// [`FreeListAllocator`] for details about the suballocation algorithm.
pub type StandardMemoryAllocator = GenericMemoryAllocator<FreeListAllocator>;
impl StandardMemoryAllocator {
/// Creates a new `StandardMemoryAllocator` with default configuration.
pub fn new_default(device: Arc<Device>) -> Self {
let MemoryProperties {
memory_types,
memory_heaps,
} = device.physical_device().memory_properties();
let mut block_sizes = vec![0; memory_types.len()];
let mut memory_type_bits = u32::MAX;
for (index, memory_type) in memory_types.iter().enumerate() {
const LARGE_HEAP_THRESHOLD: DeviceSize = 1024 * 1024 * 1024;
let heap_size = memory_heaps[memory_type.heap_index as usize].size;
block_sizes[index] = if heap_size >= LARGE_HEAP_THRESHOLD {
256 * 1024 * 1024
} else {
64 * 1024 * 1024
};
if memory_type.property_flags.intersects(
MemoryPropertyFlags::LAZILY_ALLOCATED
| MemoryPropertyFlags::PROTECTED
| MemoryPropertyFlags::DEVICE_COHERENT
| MemoryPropertyFlags::RDMA_CAPABLE,
) {
// VUID-VkMemoryAllocateInfo-memoryTypeIndex-01872
// VUID-vkAllocateMemory-deviceCoherentMemory-02790
// Lazily allocated memory would just cause problems for suballocation in general.
memory_type_bits &= !(1 << index);
}
}
let create_info = GenericMemoryAllocatorCreateInfo {
block_sizes: &block_sizes,
memory_type_bits,
..Default::default()
};
Self::new(device, create_info)
}
}
/// A generic implementation of a [memory allocator].
///
/// The allocator keeps a pool of [`DeviceMemory`] blocks for each memory type and uses the type
/// parameter `S` to [suballocate] these blocks. You can also configure the sizes of these blocks.
/// This means that you can have as many `GenericMemoryAllocator`s as you you want for different
/// needs, or for performance reasons, as long as the block sizes are configured properly so that
/// too much memory isn't wasted.
///
/// See also [the `MemoryAllocator` implementation].
///
/// # Mapping behavior
///
/// Every time a new `DeviceMemory` block is allocated, it is mapped in full automatically as long
/// as it resides in host-visible memory. It remains mapped until it is dropped, which only happens
/// if the allocator is dropped. In other words, all eligible blocks are persistently mapped, so
/// you don't need to worry about whether or not your host-visible allocations are host-accessible.
///
/// # `DeviceMemory` allocation
///
/// If an allocation is created with the [`MemoryAllocatePreference::Unknown`] option, and the
/// allocator deems the allocation too big for suballocation (larger than half the block size), or
/// the implementation prefers or requires a dedicated allocation, then that allocation is made a
/// dedicated allocation. Using [`MemoryAllocatePreference::NeverAllocate`], a dedicated allocation
/// is never created, even if the allocation is larger than the block size or a dedicated
/// allocation is required. In such a case an error is returned instead. Using
/// [`MemoryAllocatePreference::AlwaysAllocate`], a dedicated allocation is always created.
///
/// In all other cases, `DeviceMemory` is only allocated if a pool runs out of memory and needs
/// another block. No `DeviceMemory` is allocated when the allocator is created, the blocks are
/// only allocated once they are needed.
///
/// [memory allocator]: MemoryAllocator
/// [suballocate]: Suballocator
/// [the `MemoryAllocator` implementation]: Self#impl-MemoryAllocator-for-GenericMemoryAllocator<S>
#[derive(Debug)]
pub struct GenericMemoryAllocator<S> {
device: InstanceOwnedDebugWrapper<Arc<Device>>,
buffer_image_granularity: DeviceAlignment,
// Each memory type has a pool of `DeviceMemory` blocks.
pools: ArrayVec<Pool<S>, MAX_MEMORY_TYPES>,
// Global mask of memory types.
memory_type_bits: u32,
dedicated_allocation: bool,
export_handle_types: ArrayVec<ExternalMemoryHandleTypes, MAX_MEMORY_TYPES>,
flags: MemoryAllocateFlags,
// How many `DeviceMemory` allocations should be allowed before restricting them.
max_allocations: u32,
}
#[derive(Debug)]
struct Pool<S> {
blocks: Mutex<Vec<Box<Block<S>>>>,
// This is cached here for faster access, so we don't need to hop through 3 pointers.
property_flags: MemoryPropertyFlags,
atom_size: DeviceAlignment,
block_size: DeviceSize,
}
impl<S> GenericMemoryAllocator<S> {
// This is a false-positive, we only use this const for static initialization.
#[allow(clippy::declare_interior_mutable_const)]
const EMPTY_POOL: Pool<S> = Pool {
blocks: Mutex::new(Vec::new()),
property_flags: MemoryPropertyFlags::empty(),
atom_size: DeviceAlignment::MIN,
block_size: 0,
};
/// Creates a new `GenericMemoryAllocator<S>` using the provided suballocator `S` for
/// suballocation of [`DeviceMemory`] blocks.
///
/// # Panics
///
/// - Panics if `create_info.block_sizes` doesn't contain as many elements as the number of
/// memory types.
/// - Panics if `create_info.export_handle_types` is non-empty and doesn't contain as many
/// elements as the number of memory types.
pub fn new(device: Arc<Device>, create_info: GenericMemoryAllocatorCreateInfo<'_>) -> Self {
let GenericMemoryAllocatorCreateInfo {
block_sizes,
memory_type_bits,
dedicated_allocation,
export_handle_types,
mut device_address,
_ne: _,
} = create_info;
let memory_types = &device.physical_device().memory_properties().memory_types;
assert_eq!(
block_sizes.len(),
memory_types.len(),
"`create_info.block_sizes` must contain as many elements as the number of memory types",
);
if !export_handle_types.is_empty() {
assert_eq!(
export_handle_types.len(),
memory_types.len(),
"`create_info.export_handle_types` must contain as many elements as the number of \
memory types if not empty",
);
}
let buffer_image_granularity = device
.physical_device()
.properties()
.buffer_image_granularity;
let memory_types = &device.physical_device().memory_properties().memory_types;
let mut pools = ArrayVec::new(memory_types.len(), [Self::EMPTY_POOL; MAX_MEMORY_TYPES]);
for (index, &MemoryType { property_flags, .. }) in memory_types.iter().enumerate() {
pools[index].property_flags = property_flags;
if property_flags.intersects(MemoryPropertyFlags::HOST_VISIBLE)
&& !property_flags.intersects(MemoryPropertyFlags::HOST_COHERENT)
{
pools[index].atom_size =
device.physical_device().properties().non_coherent_atom_size;
}
pools[index].block_size = block_sizes[index];
}
let export_handle_types = {
let mut types = ArrayVec::new(
export_handle_types.len(),
[ExternalMemoryHandleTypes::empty(); MAX_MEMORY_TYPES],
);
types.copy_from_slice(export_handle_types);
types
};
// VUID-VkMemoryAllocateInfo-flags-03331
device_address &= device.enabled_features().buffer_device_address
&& !device.enabled_extensions().ext_buffer_device_address;
// Providers of `VkMemoryAllocateFlags`
device_address &=
device.api_version() >= Version::V1_1 || device.enabled_extensions().khr_device_group;
let flags = if device_address {
MemoryAllocateFlags::DEVICE_ADDRESS
} else {
MemoryAllocateFlags::empty()
};
let max_memory_allocation_count = device
.physical_device()
.properties()
.max_memory_allocation_count;
let max_allocations = max_memory_allocation_count / 4 * 3;
GenericMemoryAllocator {
device: InstanceOwnedDebugWrapper(device),
buffer_image_granularity,
pools,
dedicated_allocation,
export_handle_types,
flags,
memory_type_bits,
max_allocations,
}
}
#[cold]
fn allocate_device_memory(
&self,
memory_type_index: u32,
allocation_size: DeviceSize,
dedicated_allocation: Option<DedicatedAllocation<'_>>,
export_handle_types: ExternalMemoryHandleTypes,
) -> Result<Arc<DeviceMemory>, Validated<VulkanError>> {
let mut memory = DeviceMemory::allocate(
self.device.clone(),
MemoryAllocateInfo {
allocation_size,
memory_type_index,
dedicated_allocation,
export_handle_types,
flags: self.flags,
..Default::default()
},
)?;
if self.pools[memory_type_index as usize]
.property_flags
.intersects(MemoryPropertyFlags::HOST_VISIBLE)
{
// SAFETY:
// - We checked that the memory is host-visible.
// - The memory can't be mapped already, because we just allocated it.
// - Mapping the whole range is always valid.
unsafe {
memory.map_unchecked(MemoryMapInfo {
offset: 0,
size: memory.allocation_size(),
_ne: crate::NonExhaustive(()),
})?;
}
}
Ok(Arc::new(memory))
}
}
unsafe impl<S: Suballocator + Send + 'static> MemoryAllocator for GenericMemoryAllocator<S> {
fn find_memory_type_index(
&self,
memory_type_bits: u32,
filter: MemoryTypeFilter,
) -> Option<u32> {
let required_flags = filter.required_flags.into();
let preferred_flags = filter.preferred_flags.into();
let not_preferred_flags = filter.not_preferred_flags.into();
self.pools
.iter()
.map(|pool| ash::vk::MemoryPropertyFlags::from(pool.property_flags))
.enumerate()
// Filter out memory types which are supported by the memory type bits and have the
// required flags set.
.filter(|&(index, flags)| {
memory_type_bits & (1 << index) != 0 && flags & required_flags == required_flags
})
// Rank memory types with more of the preferred flags higher, and ones with more of the
// not preferred flags lower.
.min_by_key(|&(_, flags)| {
(!flags & preferred_flags).as_raw().count_ones()
+ (flags & not_preferred_flags).as_raw().count_ones()
})
.map(|(index, _)| index as u32)
}
/// Allocates memory from a specific memory type.
///
/// # Arguments
///
/// - `memory_type_index` - The index of the memory type to allocate from.
///
/// - `layout` - The layout of the allocation.
///
/// - `allocation_type` - The type of resources that can be bound to the allocation.
///
/// - `never_allocate` - If `true` then the allocator should never allocate `DeviceMemory`,
/// instead only suballocate from existing blocks.
///
/// # Panics
///
/// - Panics if `memory_type_index` is not less than the number of available memory types.
///
/// # Errors
///
/// - Returns [`AllocateDeviceMemory`] if allocating a new block failed.
/// - Returns [`OutOfPoolMemory`] if `never_allocate` is `true` and the pool doesn't have
/// enough free space.
/// - Returns [`BlockSizeExceeded`] if `create_info.layout.size()` is greater than the block
/// size corresponding to the heap that the memory type corresponding to `memory_type_index`
/// resides in.
///
/// [`AllocateDeviceMemory`]: MemoryAllocatorError::AllocateDeviceMemory
/// [`OutOfPoolMemory`]: MemoryAllocatorError::OutOfPoolMemory
/// [`BlockSizeExceeded`]: MemoryAllocatorError::BlockSizeExceeded
fn allocate_from_type(
&self,
memory_type_index: u32,
mut layout: DeviceLayout,
allocation_type: AllocationType,
never_allocate: bool,
) -> Result<MemoryAlloc, MemoryAllocatorError> {
let size = layout.size();
let pool = &self.pools[memory_type_index as usize];
if size > pool.block_size {
return Err(MemoryAllocatorError::BlockSizeExceeded);
}
layout = layout.align_to(pool.atom_size).unwrap();
let mut blocks = pool.blocks.lock();
// TODO: Incremental sorting
blocks.sort_by_key(|block| block.free_size());
let (Ok(idx) | Err(idx)) = blocks.binary_search_by_key(&size, |block| block.free_size());
for block in &mut blocks[idx..] {
if let Ok(allocation) =
block.allocate(layout, allocation_type, self.buffer_image_granularity)
{
return Ok(allocation);
}
}
if never_allocate {
return Err(MemoryAllocatorError::OutOfPoolMemory);
}
// The pool doesn't have enough real estate, so we need a new block.
let block = {
let export_handle_types = if !self.export_handle_types.is_empty() {
self.export_handle_types[memory_type_index as usize]
} else {
ExternalMemoryHandleTypes::empty()
};
let mut i = 0;
loop {
let allocation_size = pool.block_size >> i;
match self.allocate_device_memory(
memory_type_index,
allocation_size,
None,
export_handle_types,
) {
Ok(device_memory) => {
break Block::new(device_memory);
}
// Retry up to 3 times, halving the allocation size each time so long as the
// resulting size is still large enough.
Err(Validated::Error(
VulkanError::OutOfHostMemory | VulkanError::OutOfDeviceMemory,
)) if i < 3 && pool.block_size >> (i + 1) >= size => {
i += 1;
}
Err(err) => return Err(MemoryAllocatorError::AllocateDeviceMemory(err)),
}
}
};
blocks.push(block);
let block = blocks.last_mut().unwrap();
match block.allocate(layout, allocation_type, self.buffer_image_granularity) {
Ok(allocation) => Ok(allocation),
// This can't happen as we always allocate a block of sufficient size.
Err(SuballocatorError::OutOfRegionMemory) => unreachable!(),
// This can't happen as the block is fresher than Febreze and we're still holding an
// exclusive lock.
Err(SuballocatorError::FragmentedRegion) => unreachable!(),
}
}
/// Allocates memory according to requirements.
///
/// # Arguments
///
/// - `requirements` - Requirements of the resource you want to allocate memory for.
///
/// If you plan to bind this memory directly to a non-sparse resource, then this must
/// correspond to the value returned by either [`RawBuffer::memory_requirements`] or
/// [`RawImage::memory_requirements`] for the respective buffer or image.
///
/// - `allocation_type` - What type of resource this allocation will be used for.
///
/// This should be [`Linear`] for buffers and linear images, and [`NonLinear`] for optimal
/// images. You can not bind memory allocated with the [`Linear`] type to optimal images or
/// bind memory allocated with the [`NonLinear`] type to buffers and linear images. You
/// should never use the [`Unknown`] type unless you have to, as that can be less memory
/// efficient.
///
/// - `dedicated_allocation` - Allows a dedicated allocation to be created.
///
/// You should always fill this field in if you are allocating memory for a non-sparse
/// resource, otherwise the allocator won't be able to create a dedicated allocation if one
/// is required or recommended.
///
/// This argument is silently ignored (treated as `None`) if the device API version is below
/// 1.1 and the [`khr_dedicated_allocation`] extension is not enabled on the device.
///
/// # Errors
///
/// - Returns [`AllocateDeviceMemory`] if allocating a new block failed.
/// - Returns [`FindMemoryType`] if finding a suitable memory type failed. This can happen if
/// the `create_info.requirements` correspond to those of an optimal image but
/// `create_info.memory_type_filter` requires host access.
/// - Returns [`OutOfPoolMemory`] if `create_info.allocate_preference` is
/// [`MemoryAllocatePreference::NeverAllocate`] and none of the pools of suitable memory
/// types have enough free space.
/// - Returns [`DedicatedAllocationRequired`] if `create_info.allocate_preference` is
/// [`MemoryAllocatePreference::NeverAllocate`] and
/// `create_info.requirements.requires_dedicated_allocation` is `true`.
/// - Returns [`BlockSizeExceeded`] if `create_info.allocate_preference` is
/// [`MemoryAllocatePreference::NeverAllocate`] and `create_info.requirements.size` is
/// greater than the block size for all heaps of suitable memory types.
///
/// [`RawBuffer::memory_requirements`]: crate::buffer::sys::RawBuffer::memory_requirements
/// [`RawImage::memory_requirements`]: crate::image::sys::RawImage::memory_requirements
/// [`Linear`]: AllocationType::Linear
/// [`NonLinear`]: AllocationType::NonLinear
/// [`Unknown`]: AllocationType::Unknown
/// [`khr_dedicated_allocation`]: crate::device::DeviceExtensions::khr_dedicated_allocation
/// [`AllocateDeviceMemory`]: MemoryAllocatorError::AllocateDeviceMemory
/// [`FindMemoryType`]: MemoryAllocatorError::FindMemoryType
/// [`OutOfPoolMemory`]: MemoryAllocatorError::OutOfPoolMemory
/// [`DedicatedAllocationRequired`]: MemoryAllocatorError::DedicatedAllocationRequired
/// [`BlockSizeExceeded`]: MemoryAllocatorError::BlockSizeExceeded
fn allocate(
&self,
requirements: MemoryRequirements,
allocation_type: AllocationType,
create_info: AllocationCreateInfo,
mut dedicated_allocation: Option<DedicatedAllocation<'_>>,
) -> Result<MemoryAlloc, MemoryAllocatorError> {
let MemoryRequirements {
layout,
mut memory_type_bits,
mut prefers_dedicated_allocation,
requires_dedicated_allocation,
} = requirements;
memory_type_bits &= self.memory_type_bits;
memory_type_bits &= create_info.memory_type_bits;
let AllocationCreateInfo {
memory_type_filter,
memory_type_bits: _,
allocate_preference,
_ne: _,
} = create_info;
let size = layout.size();
let mut memory_type_index = self
.find_memory_type_index(memory_type_bits, memory_type_filter)
.ok_or(MemoryAllocatorError::FindMemoryType)?;
if !self.dedicated_allocation && !requires_dedicated_allocation {
dedicated_allocation = None;
}
let export_handle_types = if self.export_handle_types.is_empty() {
ExternalMemoryHandleTypes::empty()
} else {
self.export_handle_types[memory_type_index as usize]
};
loop {
let pool = &self.pools[memory_type_index as usize];
let res = match allocate_preference {
MemoryAllocatePreference::Unknown => {
// VUID-vkBindBufferMemory-buffer-01444
// VUID-vkBindImageMemory-image-01445
if requires_dedicated_allocation {
self.allocate_dedicated(
memory_type_index,
size,
dedicated_allocation,
export_handle_types,
)
} else {
if size > pool.block_size / 2 {
prefers_dedicated_allocation = true;
}
if self.device.allocation_count() > self.max_allocations
&& size <= pool.block_size
{
prefers_dedicated_allocation = false;
}
if prefers_dedicated_allocation {
self.allocate_dedicated(
memory_type_index,
size,
dedicated_allocation,
export_handle_types,
)
// Fall back to suballocation.
.or_else(|err| {
self.allocate_from_type(
memory_type_index,
layout,
allocation_type,
true, // A dedicated allocation already failed.
)
.map_err(|_| err)
})
} else {
self.allocate_from_type(
memory_type_index,
layout,
allocation_type,
false,
)
// Fall back to dedicated allocation. It is possible that the 1/8
// block size tried was greater than the allocation size, so
// there's hope.
.or_else(|_| {
self.allocate_dedicated(
memory_type_index,
size,
dedicated_allocation,
export_handle_types,
)
})
}
}
}
MemoryAllocatePreference::NeverAllocate => {
if requires_dedicated_allocation {
return Err(MemoryAllocatorError::DedicatedAllocationRequired);
}
self.allocate_from_type(memory_type_index, layout, allocation_type, true)
}
MemoryAllocatePreference::AlwaysAllocate => self.allocate_dedicated(
memory_type_index,
size,
dedicated_allocation,
export_handle_types,
),
};
match res {
Ok(allocation) => return Ok(allocation),
// Try a different memory type.
Err(err) => {
memory_type_bits &= !(1 << memory_type_index);
memory_type_index = self
.find_memory_type_index(memory_type_bits, memory_type_filter)
.ok_or(err)?;
}
}
}
}
#[cold]
fn allocate_dedicated(
&self,
memory_type_index: u32,
allocation_size: DeviceSize,
dedicated_allocation: Option<DedicatedAllocation<'_>>,
export_handle_types: ExternalMemoryHandleTypes,
) -> Result<MemoryAlloc, MemoryAllocatorError> {
let device_memory = self
.allocate_device_memory(
memory_type_index,
allocation_size,
dedicated_allocation,
export_handle_types,
)
.map_err(MemoryAllocatorError::AllocateDeviceMemory)?;
Ok(MemoryAlloc {
device_memory,
suballocation: None,
allocation_handle: AllocationHandle::null(),
})
}
unsafe fn deallocate(&self, allocation: MemoryAlloc) {
if let Some(suballocation) = allocation.suballocation {
let memory_type_index = allocation.device_memory.memory_type_index();
let pool = self.pools[memory_type_index as usize].blocks.lock();
let block_ptr = allocation.allocation_handle.0 as *mut Block<S>;
// TODO: Maybe do a similar check for dedicated blocks.
debug_assert!(
pool.iter()
.any(|block| &**block as *const Block<S> == block_ptr),
"attempted to deallocate a memory block that does not belong to this allocator",
);
// SAFETY: The caller must guarantee that `allocation` refers to one allocated by
// `self`, therefore `block_ptr` must be the same one we gave out on allocation. We
// know that this pointer must be valid, because all blocks are boxed and pinned in
// memory and because a block isn't dropped until the allocator itself is dropped, at
// which point it would be impossible to call this method. We also know that it must be
// valid to create a reference to the block, because we locked the pool it belongs to.
let block = &mut *block_ptr;
// SAFETY: The caller must guarantee that `allocation` refers to a currently allocated
// allocation of `self`.
block.deallocate(suballocation);
drop(pool);
}
}
}
unsafe impl<T: MemoryAllocator> MemoryAllocator for Arc<T> {
fn find_memory_type_index(
&self,
memory_type_bits: u32,
filter: MemoryTypeFilter,
) -> Option<u32> {
(**self).find_memory_type_index(memory_type_bits, filter)
}
fn allocate_from_type(
&self,
memory_type_index: u32,
layout: DeviceLayout,
allocation_type: AllocationType,
never_allocate: bool,
) -> Result<MemoryAlloc, MemoryAllocatorError> {
(**self).allocate_from_type(memory_type_index, layout, allocation_type, never_allocate)
}
fn allocate(
&self,
requirements: MemoryRequirements,
allocation_type: AllocationType,
create_info: AllocationCreateInfo,
dedicated_allocation: Option<DedicatedAllocation<'_>>,
) -> Result<MemoryAlloc, MemoryAllocatorError> {
(**self).allocate(
requirements,
allocation_type,
create_info,
dedicated_allocation,
)
}
fn allocate_dedicated(
&self,
memory_type_index: u32,
allocation_size: DeviceSize,
dedicated_allocation: Option<DedicatedAllocation<'_>>,
export_handle_types: ExternalMemoryHandleTypes,
) -> Result<MemoryAlloc, MemoryAllocatorError> {
(**self).allocate_dedicated(
memory_type_index,
allocation_size,
dedicated_allocation,
export_handle_types,
)
}
unsafe fn deallocate(&self, allocation: MemoryAlloc) {
(**self).deallocate(allocation)
}
}
unsafe impl<S> DeviceOwned for GenericMemoryAllocator<S> {
fn device(&self) -> &Arc<Device> {
&self.device
}
}
#[derive(Debug)]
struct Block<S> {
device_memory: Arc<DeviceMemory>,
suballocator: S,
allocation_count: usize,
}
impl<S: Suballocator> Block<S> {
fn new(device_memory: Arc<DeviceMemory>) -> Box<Self> {
let suballocator = S::new(
Region::new(0, device_memory.allocation_size())
.expect("we somehow managed to allocate more than `DeviceLayout::MAX_SIZE` bytes"),
);
Box::new(Block {
device_memory,
suballocator,
allocation_count: 0,
})
}
fn allocate(
&mut self,
layout: DeviceLayout,
allocation_type: AllocationType,
buffer_image_granularity: DeviceAlignment,
) -> Result<MemoryAlloc, SuballocatorError> {
let suballocation =
self.suballocator
.allocate(layout, allocation_type, buffer_image_granularity)?;
self.allocation_count += 1;
Ok(MemoryAlloc {
device_memory: self.device_memory.clone(),
suballocation: Some(suballocation),
allocation_handle: AllocationHandle::from_ptr(self as *mut Block<S> as _),
})
}
unsafe fn deallocate(&mut self, suballocation: Suballocation) {
self.suballocator.deallocate(suballocation);
self.allocation_count -= 1;
// For bump allocators, reset the free-start once there are no remaining allocations.
if self.allocation_count == 0 {
self.suballocator.cleanup();
}
}
fn free_size(&self) -> DeviceSize {
self.suballocator.free_size()
}
}
/// Parameters to create a new [`GenericMemoryAllocator`].
#[derive(Clone, Debug)]
pub struct GenericMemoryAllocatorCreateInfo<'a> {
/// Lets you configure the block sizes for each memory type.
///
/// Must contain one entry for each memory type. The allocator keeps a pool of [`DeviceMemory`]
/// blocks for each memory type, and every time a new block is allocated, the block size
/// corresponding to the memory type index is looked up here and used for the allocation.
///
/// The block size is going to be the maximum size of a `DeviceMemory` block that is tried. If
/// allocating a block with the size fails, the allocator tries 1/2, 1/4 and 1/8 of the block
/// size in that order until one succeeds, else a dedicated allocation is attempted for the
/// allocation. If an allocation is created with a size greater than half the block size it is
/// always made a dedicated allocation. All of this doesn't apply when using
/// [`MemoryAllocatePreference::NeverAllocate`] however.
///
/// The default value is `&[]`, which must be overridden.
pub block_sizes: &'a [DeviceSize],
/// Lets you configure the allocator's global mask of memory type indices. Only the memory type
/// indices that have a corresponding bit at the same index set will be allocated from when
/// calling [`allocate`], otherwise [`MemoryAllocatorError::FindMemoryType`] is returned.
///
/// You may use this to disallow problematic memory types, for instance ones with the
/// [`PROTECTED`] flag, or any other flags you don't want.
///
/// The default value is [`u32::MAX`].
///
/// [`allocate`]: struct.GenericMemoryAllocator.html#method.allocate
/// [`PROTECTED`]: MemoryPropertyFlags::DEVICE_COHERENT
pub memory_type_bits: u32,
/// Whether the allocator should use the dedicated allocation APIs.
///
/// This means that when the allocator decides that an allocation should not be suballocated,
/// but rather have its own block of [`DeviceMemory`], that that allocation will be made a
/// dedicated allocation. Otherwise they are still given their own block of device memory, just
/// that that block won't be [dedicated] to a particular resource.
///
/// Dedicated allocations are an optimization which may result in better performance, so there
/// really is no reason to disable this option, unless the restrictions that they bring with
/// them are a problem. Namely, a dedicated allocation must only be used for the resource it
/// was created for. Meaning that reusing the memory for something else is not possible,
/// suballocating it is not possible, and aliasing it is also not possible.
///
/// This option is silently ignored (treated as `false`) if the device API version is below 1.1
/// and the [`khr_dedicated_allocation`] extension is not enabled on the device.
///
/// The default value is `true`.
///
/// [dedicated]: DeviceMemory::is_dedicated
/// [`khr_dedicated_allocation`]: crate::device::DeviceExtensions::khr_dedicated_allocation
pub dedicated_allocation: bool,
/// Lets you configure the external memory handle types that the [`DeviceMemory`] blocks will
/// be allocated with.
///
/// Must be either empty or contain one element for each memory type. When `DeviceMemory` is
/// allocated, the external handle types corresponding to the memory type index are looked up
/// here and used for the allocation.
///
/// The default value is `&[]`.
pub export_handle_types: &'a [ExternalMemoryHandleTypes],
/// Whether the allocator should allocate the [`DeviceMemory`] blocks with the
/// [`DEVICE_ADDRESS`] flag set.
///
/// This is required if you want to allocate memory for buffers that have the
/// [`SHADER_DEVICE_ADDRESS`] usage set. For this option too, there is no reason to disable it.
///
/// This option is silently ignored (treated as `false`) if the [`buffer_device_address`]
/// feature is not enabled on the device or if the [`ext_buffer_device_address`] extension is
/// enabled on the device. It is also ignored if the device API version is below 1.1 and the
/// [`khr_device_group`] extension is not enabled on the device.
///
/// The default value is `true`.
///
/// [`DEVICE_ADDRESS`]: MemoryAllocateFlags::DEVICE_ADDRESS
/// [`SHADER_DEVICE_ADDRESS`]: crate::buffer::BufferUsage::SHADER_DEVICE_ADDRESS
/// [`buffer_device_address`]: crate::device::Features::buffer_device_address
/// [`ext_buffer_device_address`]: crate::device::DeviceExtensions::ext_buffer_device_address
/// [`khr_device_group`]: crate::device::DeviceExtensions::khr_device_group
pub device_address: bool,
pub _ne: crate::NonExhaustive,
}
impl Default for GenericMemoryAllocatorCreateInfo<'_> {
#[inline]
fn default() -> Self {
GenericMemoryAllocatorCreateInfo {
block_sizes: &[],
memory_type_bits: u32::MAX,
dedicated_allocation: true,
export_handle_types: &[],
device_address: true,
_ne: crate::NonExhaustive(()),
}
}
}
/// > **Note**: Returns `0` on overflow.
#[inline(always)]
pub(crate) const fn align_up(val: DeviceSize, alignment: DeviceAlignment) -> DeviceSize {
align_down(val.wrapping_add(alignment.as_devicesize() - 1), alignment)
}
#[inline(always)]
pub(crate) const fn align_down(val: DeviceSize, alignment: DeviceAlignment) -> DeviceSize {
val & !(alignment.as_devicesize() - 1)
}
mod array_vec {
use std::ops::{Deref, DerefMut};
/// Minimal implementation of an `ArrayVec`. Useful when a `Vec` is needed but there is a known
/// limit on the number of elements, so that it can occupy real estate on the stack.
#[derive(Clone, Copy, Debug)]
pub(super) struct ArrayVec<T, const N: usize> {
len: usize,
data: [T; N],
}
impl<T, const N: usize> ArrayVec<T, N> {
pub fn new(len: usize, data: [T; N]) -> Self {
assert!(len <= N);
ArrayVec { len, data }
}
}
impl<T, const N: usize> Deref for ArrayVec<T, N> {
type Target = [T];
fn deref(&self) -> &Self::Target {
// SAFETY: `self.len <= N`.
unsafe { self.data.get_unchecked(0..self.len) }
}
}
impl<T, const N: usize> DerefMut for ArrayVec<T, N> {
fn deref_mut(&mut self) -> &mut Self::Target {
// SAFETY: `self.len <= N`.
unsafe { self.data.get_unchecked_mut(0..self.len) }
}
}
}