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
|
/*
* Copyright © 2020 Intel Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#ifdef __linux__
#include <linux/limits.h>
#endif
#include <signal.h>
#include <sys/ioctl.h>
#include <sys/time.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <pthread.h>
#include "i915/gem_create.h"
#include "intel_reg.h"
#include "drmtest.h"
#include "ioctl_wrappers.h"
#include "igt_aux.h"
#include "igt_dummyload.h"
#include "igt_gt.h"
#include "igt_params.h"
#include "igt_sysfs.h"
#include "intel_chipset.h"
#include "igt_collection.h"
#include "igt_device.h"
#include "gem_mman.h"
#include "i915/intel_memory_region.h"
#define i915_query_items(fd, items, n_items) do { \
igt_assert_eq(__i915_query_items(fd, items, n_items), 0); \
errno = 0; \
} while (0)
#define i915_query_items_err(fd, items, n_items, err) do { \
igt_assert_eq(__i915_query_items(fd, items, n_items), -err); \
} while (0)
static int
__i915_query(int fd, struct drm_i915_query *q)
{
if (igt_ioctl(fd, DRM_IOCTL_I915_QUERY, q))
return -errno;
return 0;
}
static int
__i915_query_items(int fd, struct drm_i915_query_item *items, uint32_t n_items)
{
struct drm_i915_query q = {
.num_items = n_items,
.items_ptr = to_user_pointer(items),
};
return __i915_query(fd, &q);
}
bool gem_has_query_support(int fd)
{
struct drm_i915_query query = {};
return __i915_query(fd, &query) == 0;
}
const char *get_memory_region_name(uint32_t region)
{
uint16_t class = MEMORY_TYPE_FROM_REGION(region);
switch (class) {
case I915_MEMORY_CLASS_SYSTEM:
return "smem";
case I915_MEMORY_CLASS_DEVICE:
return "lmem";
}
igt_assert_f(false, "Unknown memory region");
}
/**
* gem_get_query_memory_regions:
* @fd: open i915 drm file descriptor
*
* This function wraps query mechanism for memory regions.
*
* Returns: Filled struct with available memory regions.
*/
struct drm_i915_query_memory_regions *gem_get_query_memory_regions(int fd)
{
struct drm_i915_query_item item;
struct drm_i915_query_memory_regions *query_info = NULL;
memset(&item, 0, sizeof(item));
item.query_id = DRM_I915_QUERY_MEMORY_REGIONS;
i915_query_items(fd, &item, 1);
/*
* Any DRM_I915_QUERY_MEMORY_REGIONS specific errors are encoded in the
* item.length, even though the ioctl might still return success.
*/
if (item.length == -ENODEV) {
/*
* If kernel supports query but not memory regions and it
* returns -ENODEV just return predefined system memory region
* only.
*/
size_t sys_qi_size = offsetof(typeof(*query_info), regions[1]);
query_info = calloc(1, sys_qi_size);
query_info->num_regions = 1;
query_info->regions[0].region.memory_class = I915_MEMORY_CLASS_SYSTEM;
goto out;
} else if (item.length < 0) {
/* Any other error are critial so no fallback is possible */
igt_critical("DRM_I915_QUERY_MEMORY_REGIONS failed with %d\n",
item.length);
goto out;
}
query_info = calloc(1, item.length);
item.data_ptr = to_user_pointer(query_info);
i915_query_items(fd, &item, 1);
out:
return query_info;
}
/**
* gem_get_lmem_region_count:
* @fd: open i915 drm file descriptor
*
* Helper function to check how many lmem regions are available on device.
*
* Returns: Number of found lmem regions.
*/
unsigned int gem_get_lmem_region_count(int fd)
{
struct drm_i915_query_memory_regions *query_info;
unsigned int lmem_regions = 0;
query_info = gem_get_query_memory_regions(fd);
if (!query_info)
goto out;
for (unsigned int i = 0; i < query_info->num_regions; i++) {
if (query_info->regions[i].region.memory_class == I915_MEMORY_CLASS_DEVICE)
lmem_regions += 1;
}
free(query_info);
out:
return lmem_regions;
}
/**
* gem_has_lmem:
* @fd: open i915 drm file descriptor
*
* Helper function to check if lmem is available on device.
*
* Returns: True if at least one lmem region was found.
*/
bool gem_has_lmem(int fd)
{
return gem_get_lmem_region_count(fd) > 0;
}
/**
* gem_has_smallbar:
* @fd: open i915 drm file descriptor
*
* Helper function to check if the device comes with small-bar.
*
* Returns: True if the device has small-bar.
*/
bool gem_has_smallbar(int fd)
{
struct drm_i915_query_memory_regions *info;
struct drm_i915_memory_region_info *rf;
bool ret = false;
info = gem_get_query_memory_regions(fd);
igt_assert(info);
for (int i = 0; i < info->num_regions; i++) {
rf = &info->regions[i];
if (rf->region.memory_class == I915_MEMORY_CLASS_DEVICE) {
if (rf->probed_size > rf->probed_cpu_visible_size) {
ret = true;
break;
}
}
}
free(info);
return ret;
}
/* A version of gem_create_in_memory_region_list which can be allowed to
fail so that the object creation can be retried */
int __gem_create_in_memory_region_list(int fd, uint32_t *handle, uint64_t *size, uint32_t flags,
const struct drm_i915_gem_memory_class_instance *mem_regions,
int num_regions)
{
struct drm_i915_gem_create_ext_memory_regions ext_regions = {
.base = { .name = I915_GEM_CREATE_EXT_MEMORY_REGIONS },
.num_regions = num_regions,
.regions = to_user_pointer(mem_regions),
};
int ret;
ret = __gem_create_ext(fd, size, flags, handle, &ext_regions.base);
if (flags && ret == -EINVAL)
ret = __gem_create_ext(fd, size, 0, handle, &ext_regions.base);
/*
* Provide fallback for stable kernels if region passed in the array
* can be system memory. In this case we get -ENODEV but still
* we're able to allocate gem bo in system memory using legacy call.
*/
if (ret == -ENODEV)
for (int i = 0; i < num_regions; i++)
if (mem_regions[i].memory_class == I915_MEMORY_CLASS_SYSTEM) {
ret = __gem_create(fd, size, handle);
break;
}
return ret;
}
/* gem_create_in_memory_region_list:
* @fd: opened i915 drm file descriptor
* @size: requested size of the buffer
* @mem_regions: memory regions array (priority list)
* @num_regions: @mem_regions length
*/
uint32_t gem_create_in_memory_region_list(int fd, uint64_t size, uint32_t flags,
const struct drm_i915_gem_memory_class_instance *mem_regions,
int num_regions)
{
uint32_t handle;
int ret = __gem_create_in_memory_region_list(fd, &handle, &size, flags,
mem_regions, num_regions);
igt_assert_eq(ret, 0);
return handle;
}
static bool __region_belongs_to_regions_type(struct drm_i915_gem_memory_class_instance region,
uint32_t *mem_regions_type,
int num_regions)
{
for (int i = 0; i < num_regions; i++)
if (mem_regions_type[i] == region.memory_class)
return true;
return false;
}
struct igt_collection *
__get_memory_region_set(struct drm_i915_query_memory_regions *regions,
uint32_t *mem_regions_type,
int num_regions)
{
struct drm_i915_gem_memory_class_instance region;
struct igt_collection *set;
int count = 0, pos = 0;
for (int i = 0; i < regions->num_regions; i++) {
region = regions->regions[i].region;
if (__region_belongs_to_regions_type(region,
mem_regions_type,
num_regions))
count++;
}
set = igt_collection_create(count);
for (int i = 0; i < regions->num_regions; i++) {
region = regions->regions[i].region;
if (__region_belongs_to_regions_type(region,
mem_regions_type,
num_regions))
igt_collection_set_value(set, pos++,
INTEL_MEMORY_REGION_ID(region.memory_class,
region.memory_instance));
}
igt_assert(count == pos);
return set;
}
/**
* memregion_dynamic_subtest_name:
* @igt_collection: memory region collection
*
* Function iterates over all memory regions inside the collection (keeped
* in the value field) and generates the name which can be used during dynamic
* subtest creation.
*
* Returns: newly allocated string, has to be freed by caller. Asserts if
* caller tries to create a name using empty collection.
*/
char *memregion_dynamic_subtest_name(struct igt_collection *set)
{
struct igt_collection_data *data;
char *name, *p;
uint32_t region, len;
igt_assert(set && set->size);
/* enough for "name%d-" * n */
len = set->size * 8;
p = name = malloc(len);
igt_assert(name);
for_each_collection_data(data, set) {
int r;
region = data->value;
if (IS_DEVICE_MEMORY_REGION(region))
r = snprintf(p, len, "%s%d-",
get_memory_region_name(region),
MEMORY_INSTANCE_FROM_REGION(region));
else
r = snprintf(p, len, "%s-",
get_memory_region_name(region));
igt_assert(r > 0);
p += r;
len -= r;
}
/* remove last '-' */
*(p - 1) = 0;
return name;
}
struct mmap_supported_region {
uint32_t region;
struct igt_list_head link;
};
/**
* get_dma_buf_mmap_supported_set:
* @i915: i915 drm file descriptor
* @set: memory regions set
*
* Function constructs set with regions which supports dma-buf mapping.
*
* Returns: set of regions which allows do dma-buf mmap or NULL otherwise.
*
* Note: set (igt_collection) need to be destroyed after use.
*/
struct igt_collection *
get_dma_buf_mmap_supported_set(int i915, struct igt_collection *set)
{
struct igt_collection *region, *supported_set = NULL;
uint32_t reg;
int dma_buf_fd;
char *ptr;
uint32_t handle, bosize = 4096;
int count = 0;
struct mmap_supported_region *mreg, *tmp;
IGT_LIST_HEAD(region_list);
for_each_combination(region, 1, set) {
reg = igt_collection_get_value(region, 0);
handle = gem_create_in_memory_regions(i915, bosize, reg);
dma_buf_fd = prime_handle_to_fd(i915, handle);
ptr = mmap(NULL, bosize, PROT_READ, MAP_SHARED, dma_buf_fd, 0);
if (ptr != MAP_FAILED) {
mreg = malloc(sizeof(*mreg));
igt_assert(mreg);
mreg->region = reg;
igt_list_add_tail(&mreg->link, ®ion_list);
count++;
}
munmap(ptr, bosize);
gem_close(i915, handle);
close(dma_buf_fd);
}
if (count) {
int i = 0;
supported_set = igt_collection_create(count);
igt_list_for_each_entry_safe(mreg, tmp, ®ion_list, link) {
igt_collection_set_value(supported_set, i++, mreg->region);
igt_list_del(&mreg->link);
free(mreg);
}
}
return supported_set;
}
/**
* intel_dump_gpu_meminfo:
* @info: pointer to drm_i915_query_memory_regions structure
*
* Outputs memory regions and their sizes.
*/
void intel_dump_gpu_meminfo(const struct drm_i915_query_memory_regions *info)
{
int i;
igt_assert(info);
igt_info("GPU meminfo:\n");
for (i = 0; i < info->num_regions; i++) {
uint32_t region = INTEL_MEMORY_REGION_ID(info->regions[i].region.memory_class,
info->regions[i].region.memory_instance);
const char *name = get_memory_region_name(region);
igt_info("- %s [%d] memory [size: 0x%llx, available: 0x%llx]\n",
name, info->regions[i].region.memory_instance,
info->regions[i].probed_size,
info->regions[i].unallocated_size);
}
}
/**
* gpu_meminfo_region_count:
* @info: pointer to drm_i915_query_memory_regions structure
* @memory_class: memory region class
*
* Returns: number of regions for type @memory_class
*/
uint32_t gpu_meminfo_region_count(const struct drm_i915_query_memory_regions *info,
uint16_t memory_class)
{
uint32_t num = 0;
int i;
igt_assert(info);
for (i = 0; i < info->num_regions; i++)
if (info->regions[i].region.memory_class == memory_class)
num++;
return num;
}
/**
* gpu_meminfo_region_total_size:
* @info: pointer to drm_i915_query_memory_regions structure
* @memory_class: memory region class
*
* Returns: total size of all regions which are type @memory_class, -1 when the
* size of at least one region is unknown
*/
uint64_t gpu_meminfo_region_total_size(const struct drm_i915_query_memory_regions *info,
uint16_t memory_class)
{
uint64_t total = 0;
int i;
igt_assert(info);
for (i = 0; i < info->num_regions; i++)
if (info->regions[i].region.memory_class == memory_class) {
if (info->regions[i].probed_size == -1)
return -1;
total += info->regions[i].probed_size;
}
return total;
}
/**
* gpu_meminfo_region_total_available:
* @info: pointer to drm_i915_query_memory_regions structure
* @memory_class: memory region class
*
* Returns: available size of all regions which are type @memory_class, -1 when
* the size of at least one region cannot be estimated
*/
uint64_t gpu_meminfo_region_total_available(const struct drm_i915_query_memory_regions *info,
uint16_t memory_class)
{
uint64_t avail = 0;
int i;
igt_assert(info);
for (i = 0; i < info->num_regions; i++)
if (info->regions[i].region.memory_class == memory_class) {
if (info->regions[i].unallocated_size == -1)
return -1;
avail += info->regions[i].unallocated_size;
}
return avail;
}
/**
* gpu_meminfo_region_size:
* @info: pointer to drm_i915_query_memory_regions structure
* @memory_class: memory region class
* @memory_instance: memory region instance
*
* Returns: available size of @memory_instance which type is @memory_class, -1
* when the size is unknown
*/
uint64_t gpu_meminfo_region_size(const struct drm_i915_query_memory_regions *info,
uint16_t memory_class,
uint16_t memory_instance)
{
int i;
igt_assert(info);
for (i = 0; i < info->num_regions; i++)
if (info->regions[i].region.memory_class == memory_class &&
info->regions[i].region.memory_instance == memory_instance)
return info->regions[i].probed_size;
return 0;
}
/**
* gpu_meminfo_region_available:
* @info: pointer to drm_i915_query_memory_regions structure
* @memory_class: memory region class
* @memory_instance: memory region instance
*
* Returns: available size of @memory_instance region which type is
* @memory_class, -1 when the size cannot be estimated
*/
uint64_t gpu_meminfo_region_available(const struct drm_i915_query_memory_regions *info,
uint16_t memory_class,
uint16_t memory_instance)
{
int i;
igt_assert(info);
for (i = 0; i < info->num_regions; i++)
if (info->regions[i].region.memory_class == memory_class &&
info->regions[i].region.memory_instance == memory_instance)
return info->regions[i].unallocated_size;
return 0;
}
#define PAGE_SIZE 4096
enum cache_entry_type {
MIN_START_OFFSET,
MIN_ALIGNMENT,
SAFE_START_OFFSET,
SAFE_ALIGNMENT,
};
struct cache_entry {
uint16_t devid;
enum cache_entry_type type;
union {
/* for MIN_START_OFFSET */
struct {
uint64_t offset;
uint32_t region;
} start;
/* for MIN_ALIGNMENT */
struct {
uint64_t alignment;
uint64_t region1;
uint64_t region2;
} minalign;
/* for SAFE_START_OFFSET */
uint64_t safe_start_offset;
/* for SAFE_ALIGNMENT */
uint64_t safe_alignment;
};
struct igt_list_head link;
};
static IGT_LIST_HEAD(cache);
static pthread_mutex_t cache_mutex = PTHREAD_MUTEX_INITIALIZER;
static struct cache_entry *find_entry_unlocked(enum cache_entry_type type,
uint16_t devid,
uint32_t region1,
uint32_t region2)
{
struct cache_entry *entry;
igt_list_for_each_entry(entry, &cache, link) {
if (entry->type != type || entry->devid != devid)
continue;
switch (entry->type) {
case MIN_START_OFFSET:
if (entry->start.region == region1)
return entry;
continue;
case MIN_ALIGNMENT:
if (entry->minalign.region1 == region1 &&
entry->minalign.region2 == region2)
return entry;
continue;
case SAFE_START_OFFSET:
case SAFE_ALIGNMENT:
return entry;
}
}
return NULL;
}
/**
* gem_detect_min_start_offset_for_region:
* @i915: drm fd
* @region: memory region
*
* Returns: minimum start offset at which kernel allows placing objects
* for memory region.
*/
uint64_t gem_detect_min_start_offset_for_region(int i915, uint32_t region)
{
struct drm_i915_gem_exec_object2 obj;
struct drm_i915_gem_execbuffer2 eb;
uint64_t start_offset = 0;
uint64_t bb_size = PAGE_SIZE;
uint32_t *batch, ctx = 0;
uint16_t devid = intel_get_drm_devid(i915);
struct cache_entry *entry, *newentry;
pthread_mutex_lock(&cache_mutex);
entry = find_entry_unlocked(MIN_START_OFFSET, devid, region, 0);
if (entry)
goto out;
pthread_mutex_unlock(&cache_mutex);
/* Use separate context if possible to avoid offset overlapping */
__gem_context_create(i915, &ctx);
memset(&obj, 0, sizeof(obj));
memset(&eb, 0, sizeof(eb));
eb.buffers_ptr = to_user_pointer(&obj);
eb.buffer_count = 1;
eb.flags = I915_EXEC_DEFAULT;
eb.rsvd1 = ctx;
igt_assert(__gem_create_in_memory_regions(i915, &obj.handle, &bb_size, region) == 0);
obj.flags = EXEC_OBJECT_PINNED;
batch = gem_mmap__device_coherent(i915, obj.handle, 0, bb_size, PROT_WRITE);
*batch = MI_BATCH_BUFFER_END;
munmap(batch, bb_size);
while (1) {
obj.offset = start_offset;
if (__gem_execbuf(i915, &eb) == 0)
break;
if (start_offset)
start_offset <<= 1;
else
start_offset = PAGE_SIZE;
if (start_offset >= 1ull << 32)
obj.flags |= EXEC_OBJECT_SUPPORTS_48B_ADDRESS;
igt_assert(start_offset <= 1ull << 48);
}
gem_close(i915, obj.handle);
if (ctx)
gem_context_destroy(i915, ctx);
newentry = malloc(sizeof(*newentry));
if (!newentry)
return start_offset;
/* Check does other thread did the job before */
pthread_mutex_lock(&cache_mutex);
entry = find_entry_unlocked(MIN_START_OFFSET, devid, region, 0);
if (entry)
goto out;
entry = newentry;
entry->devid = devid;
entry->type = MIN_START_OFFSET;
entry->start.offset = start_offset;
entry->start.region = region;
igt_list_add(&entry->link, &cache);
out:
pthread_mutex_unlock(&cache_mutex);
return entry->start.offset;
}
/**
* gem_detect_safe_start_offset:
* @i915: drm fd
*
* Returns: finds start offset which can be used as first one regardless
* memory region. Useful if for some reason some regions don't allow
* starting from 0x0 offset.
*/
uint64_t gem_detect_safe_start_offset(int i915)
{
struct drm_i915_query_memory_regions *query_info;
struct igt_collection *regions, *set;
uint32_t region;
uint64_t offset = 0;
uint16_t devid = intel_get_drm_devid(i915);
struct cache_entry *entry, *newentry;
pthread_mutex_lock(&cache_mutex);
entry = find_entry_unlocked(SAFE_START_OFFSET, devid, 0, 0);
if (entry)
goto out;
pthread_mutex_unlock(&cache_mutex);
query_info = gem_get_query_memory_regions(i915);
igt_assert(query_info);
set = get_memory_region_set(query_info,
I915_SYSTEM_MEMORY,
I915_DEVICE_MEMORY);
for_each_combination(regions, 1, set) {
region = igt_collection_get_value(regions, 0);
offset = max(offset,
gem_detect_min_start_offset_for_region(i915, region));
}
free(query_info);
igt_collection_destroy(set);
newentry = malloc(sizeof(*newentry));
if (!newentry)
return offset;
pthread_mutex_lock(&cache_mutex);
entry = find_entry_unlocked(SAFE_START_OFFSET, devid, 0, 0);
if (entry)
goto out;
entry = newentry;
entry->devid = devid;
entry->type = SAFE_START_OFFSET;
entry->safe_start_offset = offset;
igt_list_add(&entry->link, &cache);
out:
pthread_mutex_unlock(&cache_mutex);
return entry->safe_start_offset;
}
/**
* gem_detect_min_alignment_for_regions:
* @i915: drm fd
* @region1: first region
* @region2: second region
*
* Returns: minimum alignment which must be used when objects from @region1 and
* @region2 are going to interact.
*/
uint64_t gem_detect_min_alignment_for_regions(int i915,
uint32_t region1,
uint32_t region2)
{
struct drm_i915_gem_exec_object2 obj[2];
struct drm_i915_gem_execbuffer2 eb;
uint64_t min_alignment = PAGE_SIZE;
uint64_t bb_size = PAGE_SIZE, obj_size = PAGE_SIZE;
uint32_t *batch, ctx = 0;
uint16_t devid = intel_get_drm_devid(i915);
struct cache_entry *entry, *newentry;
pthread_mutex_lock(&cache_mutex);
entry = find_entry_unlocked(MIN_ALIGNMENT, devid, region1, region2);
if (entry)
goto out;
pthread_mutex_unlock(&cache_mutex);
/* Use separate context if possible to avoid offset overlapping */
__gem_context_create(i915, &ctx);
memset(obj, 0, sizeof(obj));
memset(&eb, 0, sizeof(eb));
/* Establish bb offset first */
eb.buffers_ptr = to_user_pointer(obj);
eb.buffer_count = ARRAY_SIZE(obj);
eb.flags = I915_EXEC_BATCH_FIRST | I915_EXEC_DEFAULT;
eb.rsvd1 = ctx;
igt_assert(__gem_create_in_memory_regions(i915, &obj[0].handle,
&bb_size, region1) == 0);
batch = gem_mmap__device_coherent(i915, obj[0].handle, 0, bb_size,
PROT_WRITE);
*batch = MI_BATCH_BUFFER_END;
munmap(batch, bb_size);
obj[0].flags = EXEC_OBJECT_PINNED;
obj[0].offset = gem_detect_min_start_offset_for_region(i915, region1);
/* Find appropriate alignment of object */
igt_assert(__gem_create_in_memory_regions(i915, &obj[1].handle,
&obj_size, region2) == 0);
obj[1].handle = gem_create_in_memory_regions(i915, PAGE_SIZE, region2);
obj[1].flags = EXEC_OBJECT_PINNED;
while (1) {
obj[1].offset = ALIGN(obj[0].offset + bb_size, min_alignment);
igt_assert(obj[1].offset <= 1ull << 32);
if (__gem_execbuf(i915, &eb) == 0)
break;
min_alignment <<= 1;
}
gem_close(i915, obj[0].handle);
gem_close(i915, obj[1].handle);
if (ctx)
gem_context_destroy(i915, ctx);
newentry = malloc(sizeof(*newentry));
if (!newentry)
return min_alignment;
pthread_mutex_lock(&cache_mutex);
entry = find_entry_unlocked(MIN_ALIGNMENT, devid, region1, region2);
if (entry)
goto out;
entry = newentry;
entry->devid = devid;
entry->type = MIN_ALIGNMENT;
entry->minalign.alignment = min_alignment;
entry->minalign.region1 = region1;
entry->minalign.region2 = region2;
igt_list_add(&entry->link, &cache);
out:
pthread_mutex_unlock(&cache_mutex);
return entry->minalign.alignment;
}
/**
* gem_detect_safe_alignment:
* @i915: drm fd
*
* Returns: safe alignment for all memory regions on @i915 device.
* Safe in this case means max() from all minimum alignments for each
* region.
*/
uint64_t gem_detect_safe_alignment(int i915)
{
struct drm_i915_query_memory_regions *query_info;
struct igt_collection *regions, *set;
uint64_t default_alignment = 0;
uint32_t region_bb, region_obj;
uint16_t devid = intel_get_drm_devid(i915);
struct cache_entry *entry, *newentry;
/* non-discrete uses 4K page size */
if (!gem_has_lmem(i915))
return PAGE_SIZE;
pthread_mutex_lock(&cache_mutex);
entry = find_entry_unlocked(SAFE_ALIGNMENT, devid, 0, 0);
if (entry)
goto out;
pthread_mutex_unlock(&cache_mutex);
query_info = gem_get_query_memory_regions(i915);
igt_assert(query_info);
set = get_memory_region_set(query_info,
I915_SYSTEM_MEMORY,
I915_DEVICE_MEMORY);
for_each_variation_r(regions, 2, set) {
uint64_t alignment;
region_bb = igt_collection_get_value(regions, 0);
region_obj = igt_collection_get_value(regions, 1);
/* We're interested in triangular matrix */
if (region_bb > region_obj)
continue;
alignment = gem_detect_min_alignment_for_regions(i915,
region_bb,
region_obj);
if (default_alignment < alignment)
default_alignment = alignment;
}
free(query_info);
igt_collection_destroy(set);
newentry = malloc(sizeof(*newentry));
if (!newentry)
return default_alignment;
/* Try again, check does we have cache updated in the meantime. */
pthread_mutex_lock(&cache_mutex);
entry = find_entry_unlocked(SAFE_ALIGNMENT, devid, 0, 0);
if (entry)
goto out;
entry = newentry;
entry->devid = devid;
entry->type = SAFE_ALIGNMENT;
entry->safe_alignment = default_alignment;
igt_list_add(&entry->link, &cache);
out:
pthread_mutex_unlock(&cache_mutex);
return entry->minalign.alignment;
}
static const char *
region_repr(const struct drm_i915_gem_memory_class_instance *ci)
{
switch (ci->memory_class) {
case I915_MEMORY_CLASS_SYSTEM:
return "smem";
case I915_MEMORY_CLASS_DEVICE:
return "lmem";
default:
return "unknown";
}
}
struct gem_memory_region *__gem_get_memory_regions(int i915)
{
struct drm_i915_query_memory_regions *info;
struct gem_memory_region *first = NULL;
info = gem_get_query_memory_regions(i915);
for (int i = 0; info && i < info->num_regions; i++) {
struct gem_memory_region *r;
r = malloc(sizeof(*r));
igt_assert(r);
r->ci = info->regions[i].region;
r->size = info->regions[i].probed_size;
r->cpu_size = info->regions[i].probed_cpu_visible_size;
if (r->size == -1ull)
r->size = igt_get_avail_ram_mb() << 20;
asprintf(&r->name, "%s%d",
region_repr(&r->ci), r->ci.memory_instance);
r->next = first;
first = r;
}
free(info);
return first;
}
struct gem_memory_region *__gem_next_memory_region(struct gem_memory_region *r)
{
struct gem_memory_region *next = r->next;
free(r->name);
free(r);
return next;
}
|