2 * Copyright © 2015 Intel Corporation
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
18 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
28 #include <sys/sysinfo.h>
32 #include "drm-uapi/drm_fourcc.h"
34 #include "anv_private.h"
35 #include "util/strtod.h"
36 #include "util/debug.h"
37 #include "util/build_id.h"
38 #include "util/disk_cache.h"
39 #include "util/mesa-sha1.h"
40 #include "util/os_file.h"
41 #include "util/u_atomic.h"
42 #include "util/u_string.h"
45 #include "common/gen_defines.h"
46 #include "compiler/glsl_types.h"
48 #include "genxml/gen7_pack.h"
50 /* This is probably far to big but it reflects the max size used for messages
51 * in OpenGLs KHR_debug.
53 #define MAX_DEBUG_MESSAGE_LENGTH 4096
56 compiler_debug_log(void *data
, const char *fmt
, ...)
58 char str
[MAX_DEBUG_MESSAGE_LENGTH
];
59 struct anv_device
*device
= (struct anv_device
*)data
;
61 if (list_empty(&device
->instance
->debug_report_callbacks
.callbacks
))
66 (void) vsnprintf(str
, MAX_DEBUG_MESSAGE_LENGTH
, fmt
, args
);
69 vk_debug_report(&device
->instance
->debug_report_callbacks
,
70 VK_DEBUG_REPORT_DEBUG_BIT_EXT
,
71 VK_DEBUG_REPORT_OBJECT_TYPE_UNKNOWN_EXT
,
76 compiler_perf_log(void *data
, const char *fmt
, ...)
81 if (unlikely(INTEL_DEBUG
& DEBUG_PERF
))
82 intel_logd_v(fmt
, args
);
88 anv_compute_heap_size(int fd
, uint64_t gtt_size
)
90 /* Query the total ram from the system */
94 uint64_t total_ram
= (uint64_t)info
.totalram
* (uint64_t)info
.mem_unit
;
96 /* We don't want to burn too much ram with the GPU. If the user has 4GiB
97 * or less, we use at most half. If they have more than 4GiB, we use 3/4.
99 uint64_t available_ram
;
100 if (total_ram
<= 4ull * 1024ull * 1024ull * 1024ull)
101 available_ram
= total_ram
/ 2;
103 available_ram
= total_ram
* 3 / 4;
105 /* We also want to leave some padding for things we allocate in the driver,
106 * so don't go over 3/4 of the GTT either.
108 uint64_t available_gtt
= gtt_size
* 3 / 4;
110 return MIN2(available_ram
, available_gtt
);
114 anv_physical_device_init_heaps(struct anv_physical_device
*device
, int fd
)
117 if (anv_gem_get_context_param(fd
, 0, I915_CONTEXT_PARAM_GTT_SIZE
,
119 /* If, for whatever reason, we can't actually get the GTT size from the
120 * kernel (too old?) fall back to the aperture size.
122 anv_perf_warn(NULL
, NULL
,
123 "Failed to get I915_CONTEXT_PARAM_GTT_SIZE: %m");
125 if (anv_gem_get_aperture(fd
, >t_size
) == -1) {
126 return vk_errorf(NULL
, NULL
, VK_ERROR_INITIALIZATION_FAILED
,
127 "failed to get aperture size: %m");
131 device
->supports_48bit_addresses
= (device
->info
.gen
>= 8) &&
132 gtt_size
> (4ULL << 30 /* GiB */);
134 uint64_t heap_size
= anv_compute_heap_size(fd
, gtt_size
);
136 if (heap_size
> (2ull << 30) && !device
->supports_48bit_addresses
) {
137 /* When running with an overridden PCI ID, we may get a GTT size from
138 * the kernel that is greater than 2 GiB but the execbuf check for 48bit
139 * address support can still fail. Just clamp the address space size to
140 * 2 GiB if we don't have 48-bit support.
142 intel_logw("%s:%d: The kernel reported a GTT size larger than 2 GiB but "
143 "not support for 48-bit addresses",
145 heap_size
= 2ull << 30;
148 if (heap_size
<= 3ull * (1ull << 30)) {
149 /* In this case, everything fits nicely into the 32-bit address space,
150 * so there's no need for supporting 48bit addresses on client-allocated
153 device
->memory
.heap_count
= 1;
154 device
->memory
.heaps
[0] = (struct anv_memory_heap
) {
155 .vma_start
= LOW_HEAP_MIN_ADDRESS
,
156 .vma_size
= LOW_HEAP_SIZE
,
158 .flags
= VK_MEMORY_HEAP_DEVICE_LOCAL_BIT
,
159 .supports_48bit_addresses
= false,
162 /* Not everything will fit nicely into a 32-bit address space. In this
163 * case we need a 64-bit heap. Advertise a small 32-bit heap and a
164 * larger 48-bit heap. If we're in this case, then we have a total heap
165 * size larger than 3GiB which most likely means they have 8 GiB of
166 * video memory and so carving off 1 GiB for the 32-bit heap should be
169 const uint64_t heap_size_32bit
= 1ull << 30;
170 const uint64_t heap_size_48bit
= heap_size
- heap_size_32bit
;
172 assert(device
->supports_48bit_addresses
);
174 device
->memory
.heap_count
= 2;
175 device
->memory
.heaps
[0] = (struct anv_memory_heap
) {
176 .vma_start
= HIGH_HEAP_MIN_ADDRESS
,
177 /* Leave the last 4GiB out of the high vma range, so that no state
178 * base address + size can overflow 48 bits. For more information see
179 * the comment about Wa32bitGeneralStateOffset in anv_allocator.c
181 .vma_size
= gtt_size
- (1ull << 32) - HIGH_HEAP_MIN_ADDRESS
,
182 .size
= heap_size_48bit
,
183 .flags
= VK_MEMORY_HEAP_DEVICE_LOCAL_BIT
,
184 .supports_48bit_addresses
= true,
186 device
->memory
.heaps
[1] = (struct anv_memory_heap
) {
187 .vma_start
= LOW_HEAP_MIN_ADDRESS
,
188 .vma_size
= LOW_HEAP_SIZE
,
189 .size
= heap_size_32bit
,
190 .flags
= VK_MEMORY_HEAP_DEVICE_LOCAL_BIT
,
191 .supports_48bit_addresses
= false,
195 uint32_t type_count
= 0;
196 for (uint32_t heap
= 0; heap
< device
->memory
.heap_count
; heap
++) {
197 uint32_t valid_buffer_usage
= ~0;
199 /* There appears to be a hardware issue in the VF cache where it only
200 * considers the bottom 32 bits of memory addresses. If you happen to
201 * have two vertex buffers which get placed exactly 4 GiB apart and use
202 * them in back-to-back draw calls, you can get collisions. In order to
203 * solve this problem, we require vertex and index buffers be bound to
204 * memory allocated out of the 32-bit heap.
206 if (device
->memory
.heaps
[heap
].supports_48bit_addresses
) {
207 valid_buffer_usage
&= ~(VK_BUFFER_USAGE_INDEX_BUFFER_BIT
|
208 VK_BUFFER_USAGE_VERTEX_BUFFER_BIT
);
211 if (device
->info
.has_llc
) {
212 /* Big core GPUs share LLC with the CPU and thus one memory type can be
213 * both cached and coherent at the same time.
215 device
->memory
.types
[type_count
++] = (struct anv_memory_type
) {
216 .propertyFlags
= VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT
|
217 VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT
|
218 VK_MEMORY_PROPERTY_HOST_COHERENT_BIT
|
219 VK_MEMORY_PROPERTY_HOST_CACHED_BIT
,
221 .valid_buffer_usage
= valid_buffer_usage
,
224 /* The spec requires that we expose a host-visible, coherent memory
225 * type, but Atom GPUs don't share LLC. Thus we offer two memory types
226 * to give the application a choice between cached, but not coherent and
227 * coherent but uncached (WC though).
229 device
->memory
.types
[type_count
++] = (struct anv_memory_type
) {
230 .propertyFlags
= VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT
|
231 VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT
|
232 VK_MEMORY_PROPERTY_HOST_COHERENT_BIT
,
234 .valid_buffer_usage
= valid_buffer_usage
,
236 device
->memory
.types
[type_count
++] = (struct anv_memory_type
) {
237 .propertyFlags
= VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT
|
238 VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT
|
239 VK_MEMORY_PROPERTY_HOST_CACHED_BIT
,
241 .valid_buffer_usage
= valid_buffer_usage
,
245 device
->memory
.type_count
= type_count
;
251 anv_physical_device_init_uuids(struct anv_physical_device
*device
)
253 const struct build_id_note
*note
=
254 build_id_find_nhdr_for_addr(anv_physical_device_init_uuids
);
256 return vk_errorf(device
->instance
, device
,
257 VK_ERROR_INITIALIZATION_FAILED
,
258 "Failed to find build-id");
261 unsigned build_id_len
= build_id_length(note
);
262 if (build_id_len
< 20) {
263 return vk_errorf(device
->instance
, device
,
264 VK_ERROR_INITIALIZATION_FAILED
,
265 "build-id too short. It needs to be a SHA");
268 memcpy(device
->driver_build_sha1
, build_id_data(note
), 20);
270 struct mesa_sha1 sha1_ctx
;
272 STATIC_ASSERT(VK_UUID_SIZE
<= sizeof(sha1
));
274 /* The pipeline cache UUID is used for determining when a pipeline cache is
275 * invalid. It needs both a driver build and the PCI ID of the device.
277 _mesa_sha1_init(&sha1_ctx
);
278 _mesa_sha1_update(&sha1_ctx
, build_id_data(note
), build_id_len
);
279 _mesa_sha1_update(&sha1_ctx
, &device
->chipset_id
,
280 sizeof(device
->chipset_id
));
281 _mesa_sha1_update(&sha1_ctx
, &device
->always_use_bindless
,
282 sizeof(device
->always_use_bindless
));
283 _mesa_sha1_update(&sha1_ctx
, &device
->has_a64_buffer_access
,
284 sizeof(device
->has_a64_buffer_access
));
285 _mesa_sha1_update(&sha1_ctx
, &device
->has_bindless_images
,
286 sizeof(device
->has_bindless_images
));
287 _mesa_sha1_update(&sha1_ctx
, &device
->has_bindless_samplers
,
288 sizeof(device
->has_bindless_samplers
));
289 _mesa_sha1_final(&sha1_ctx
, sha1
);
290 memcpy(device
->pipeline_cache_uuid
, sha1
, VK_UUID_SIZE
);
292 /* The driver UUID is used for determining sharability of images and memory
293 * between two Vulkan instances in separate processes. People who want to
294 * share memory need to also check the device UUID (below) so all this
295 * needs to be is the build-id.
297 memcpy(device
->driver_uuid
, build_id_data(note
), VK_UUID_SIZE
);
299 /* The device UUID uniquely identifies the given device within the machine.
300 * Since we never have more than one device, this doesn't need to be a real
301 * UUID. However, on the off-chance that someone tries to use this to
302 * cache pre-tiled images or something of the like, we use the PCI ID and
303 * some bits of ISL info to ensure that this is safe.
305 _mesa_sha1_init(&sha1_ctx
);
306 _mesa_sha1_update(&sha1_ctx
, &device
->chipset_id
,
307 sizeof(device
->chipset_id
));
308 _mesa_sha1_update(&sha1_ctx
, &device
->isl_dev
.has_bit6_swizzling
,
309 sizeof(device
->isl_dev
.has_bit6_swizzling
));
310 _mesa_sha1_final(&sha1_ctx
, sha1
);
311 memcpy(device
->device_uuid
, sha1
, VK_UUID_SIZE
);
317 anv_physical_device_init_disk_cache(struct anv_physical_device
*device
)
319 #ifdef ENABLE_SHADER_CACHE
321 ASSERTED
int len
= snprintf(renderer
, sizeof(renderer
), "anv_%04x",
323 assert(len
== sizeof(renderer
) - 2);
326 _mesa_sha1_format(timestamp
, device
->driver_build_sha1
);
328 const uint64_t driver_flags
=
329 brw_get_compiler_config_value(device
->compiler
);
330 device
->disk_cache
= disk_cache_create(renderer
, timestamp
, driver_flags
);
332 device
->disk_cache
= NULL
;
337 anv_physical_device_free_disk_cache(struct anv_physical_device
*device
)
339 #ifdef ENABLE_SHADER_CACHE
340 if (device
->disk_cache
)
341 disk_cache_destroy(device
->disk_cache
);
343 assert(device
->disk_cache
== NULL
);
348 get_available_system_memory()
350 char *meminfo
= os_read_file("/proc/meminfo");
354 char *str
= strstr(meminfo
, "MemAvailable:");
360 uint64_t kb_mem_available
;
361 if (sscanf(str
, "MemAvailable: %" PRIx64
, &kb_mem_available
) == 1) {
363 return kb_mem_available
<< 10;
371 anv_physical_device_init(struct anv_physical_device
*device
,
372 struct anv_instance
*instance
,
373 drmDevicePtr drm_device
)
375 const char *primary_path
= drm_device
->nodes
[DRM_NODE_PRIMARY
];
376 const char *path
= drm_device
->nodes
[DRM_NODE_RENDER
];
381 brw_process_intel_debug_variable();
383 fd
= open(path
, O_RDWR
| O_CLOEXEC
);
385 return vk_error(VK_ERROR_INCOMPATIBLE_DRIVER
);
387 device
->_loader_data
.loaderMagic
= ICD_LOADER_MAGIC
;
388 device
->instance
= instance
;
390 assert(strlen(path
) < ARRAY_SIZE(device
->path
));
391 snprintf(device
->path
, ARRAY_SIZE(device
->path
), "%s", path
);
393 if (!gen_get_device_info_from_fd(fd
, &device
->info
)) {
394 result
= vk_error(VK_ERROR_INCOMPATIBLE_DRIVER
);
397 device
->chipset_id
= device
->info
.chipset_id
;
398 device
->no_hw
= device
->info
.no_hw
;
400 if (getenv("INTEL_NO_HW") != NULL
)
401 device
->no_hw
= true;
403 device
->pci_info
.domain
= drm_device
->businfo
.pci
->domain
;
404 device
->pci_info
.bus
= drm_device
->businfo
.pci
->bus
;
405 device
->pci_info
.device
= drm_device
->businfo
.pci
->dev
;
406 device
->pci_info
.function
= drm_device
->businfo
.pci
->func
;
408 device
->name
= gen_get_device_name(device
->chipset_id
);
410 if (device
->info
.is_haswell
) {
411 intel_logw("Haswell Vulkan support is incomplete");
412 } else if (device
->info
.gen
== 7 && !device
->info
.is_baytrail
) {
413 intel_logw("Ivy Bridge Vulkan support is incomplete");
414 } else if (device
->info
.gen
== 7 && device
->info
.is_baytrail
) {
415 intel_logw("Bay Trail Vulkan support is incomplete");
416 } else if (device
->info
.gen
>= 8 && device
->info
.gen
<= 11) {
417 /* Gen8-11 fully supported */
419 result
= vk_errorf(device
->instance
, device
,
420 VK_ERROR_INCOMPATIBLE_DRIVER
,
421 "Vulkan not yet supported on %s", device
->name
);
425 device
->cmd_parser_version
= -1;
426 if (device
->info
.gen
== 7) {
427 device
->cmd_parser_version
=
428 anv_gem_get_param(fd
, I915_PARAM_CMD_PARSER_VERSION
);
429 if (device
->cmd_parser_version
== -1) {
430 result
= vk_errorf(device
->instance
, device
,
431 VK_ERROR_INITIALIZATION_FAILED
,
432 "failed to get command parser version");
437 if (!anv_gem_get_param(fd
, I915_PARAM_HAS_WAIT_TIMEOUT
)) {
438 result
= vk_errorf(device
->instance
, device
,
439 VK_ERROR_INITIALIZATION_FAILED
,
440 "kernel missing gem wait");
444 if (!anv_gem_get_param(fd
, I915_PARAM_HAS_EXECBUF2
)) {
445 result
= vk_errorf(device
->instance
, device
,
446 VK_ERROR_INITIALIZATION_FAILED
,
447 "kernel missing execbuf2");
451 if (!device
->info
.has_llc
&&
452 anv_gem_get_param(fd
, I915_PARAM_MMAP_VERSION
) < 1) {
453 result
= vk_errorf(device
->instance
, device
,
454 VK_ERROR_INITIALIZATION_FAILED
,
455 "kernel missing wc mmap");
459 result
= anv_physical_device_init_heaps(device
, fd
);
460 if (result
!= VK_SUCCESS
)
463 device
->has_exec_async
= anv_gem_get_param(fd
, I915_PARAM_HAS_EXEC_ASYNC
);
464 device
->has_exec_capture
= anv_gem_get_param(fd
, I915_PARAM_HAS_EXEC_CAPTURE
);
465 device
->has_exec_fence
= anv_gem_get_param(fd
, I915_PARAM_HAS_EXEC_FENCE
);
466 device
->has_syncobj
= anv_gem_get_param(fd
, I915_PARAM_HAS_EXEC_FENCE_ARRAY
);
467 device
->has_syncobj_wait
= device
->has_syncobj
&&
468 anv_gem_supports_syncobj_wait(fd
);
469 device
->has_context_priority
= anv_gem_has_context_priority(fd
);
471 device
->use_softpin
= anv_gem_get_param(fd
, I915_PARAM_HAS_EXEC_SOFTPIN
)
472 && device
->supports_48bit_addresses
;
474 device
->has_context_isolation
=
475 anv_gem_get_param(fd
, I915_PARAM_HAS_CONTEXT_ISOLATION
);
477 device
->always_use_bindless
=
478 env_var_as_boolean("ANV_ALWAYS_BINDLESS", false);
480 /* We first got the A64 messages on broadwell and we can only use them if
481 * we can pass addresses directly into the shader which requires softpin.
483 device
->has_a64_buffer_access
= device
->info
.gen
>= 8 &&
486 /* We first get bindless image access on Skylake and we can only really do
487 * it if we don't have any relocations so we need softpin.
489 device
->has_bindless_images
= device
->info
.gen
>= 9 &&
492 /* We've had bindless samplers since Ivy Bridge (forever in Vulkan terms)
493 * because it's just a matter of setting the sampler address in the sample
494 * message header. However, we've not bothered to wire it up for vec4 so
495 * we leave it disabled on gen7.
497 device
->has_bindless_samplers
= device
->info
.gen
>= 8;
499 device
->has_mem_available
= get_available_system_memory() != 0;
501 /* Starting with Gen10, the timestamp frequency of the command streamer may
502 * vary from one part to another. We can query the value from the kernel.
504 if (device
->info
.gen
>= 10) {
505 int timestamp_frequency
=
506 anv_gem_get_param(fd
, I915_PARAM_CS_TIMESTAMP_FREQUENCY
);
508 if (timestamp_frequency
< 0)
509 intel_logw("Kernel 4.16-rc1+ required to properly query CS timestamp frequency");
511 device
->info
.timestamp_frequency
= timestamp_frequency
;
514 /* GENs prior to 8 do not support EU/Subslice info */
515 if (device
->info
.gen
>= 8) {
516 device
->subslice_total
= anv_gem_get_param(fd
, I915_PARAM_SUBSLICE_TOTAL
);
517 device
->eu_total
= anv_gem_get_param(fd
, I915_PARAM_EU_TOTAL
);
519 /* Without this information, we cannot get the right Braswell
520 * brandstrings, and we have to use conservative numbers for GPGPU on
521 * many platforms, but otherwise, things will just work.
523 if (device
->subslice_total
< 1 || device
->eu_total
< 1) {
524 intel_logw("Kernel 4.1 required to properly query GPU properties");
526 } else if (device
->info
.gen
== 7) {
527 device
->subslice_total
= 1 << (device
->info
.gt
- 1);
530 if (device
->info
.is_cherryview
&&
531 device
->subslice_total
> 0 && device
->eu_total
> 0) {
532 /* Logical CS threads = EUs per subslice * num threads per EU */
533 uint32_t max_cs_threads
=
534 device
->eu_total
/ device
->subslice_total
* device
->info
.num_thread_per_eu
;
536 /* Fuse configurations may give more threads than expected, never less. */
537 if (max_cs_threads
> device
->info
.max_cs_threads
)
538 device
->info
.max_cs_threads
= max_cs_threads
;
541 device
->compiler
= brw_compiler_create(NULL
, &device
->info
);
542 if (device
->compiler
== NULL
) {
543 result
= vk_error(VK_ERROR_OUT_OF_HOST_MEMORY
);
546 device
->compiler
->shader_debug_log
= compiler_debug_log
;
547 device
->compiler
->shader_perf_log
= compiler_perf_log
;
548 device
->compiler
->supports_pull_constants
= false;
549 device
->compiler
->constant_buffer_0_is_relative
=
550 device
->info
.gen
< 8 || !device
->has_context_isolation
;
551 device
->compiler
->supports_shader_constants
= true;
553 /* Broadwell PRM says:
555 * "Before Gen8, there was a historical configuration control field to
556 * swizzle address bit[6] for in X/Y tiling modes. This was set in three
557 * different places: TILECTL[1:0], ARB_MODE[5:4], and
558 * DISP_ARB_CTL[14:13].
560 * For Gen8 and subsequent generations, the swizzle fields are all
561 * reserved, and the CPU's memory controller performs all address
562 * swizzling modifications."
565 device
->info
.gen
< 8 && anv_gem_get_bit6_swizzle(fd
, I915_TILING_X
);
567 isl_device_init(&device
->isl_dev
, &device
->info
, swizzled
);
569 result
= anv_physical_device_init_uuids(device
);
570 if (result
!= VK_SUCCESS
)
573 anv_physical_device_init_disk_cache(device
);
575 if (instance
->enabled_extensions
.KHR_display
) {
576 master_fd
= open(primary_path
, O_RDWR
| O_CLOEXEC
);
577 if (master_fd
>= 0) {
578 /* prod the device with a GETPARAM call which will fail if
579 * we don't have permission to even render on this device
581 if (anv_gem_get_param(master_fd
, I915_PARAM_CHIPSET_ID
) == 0) {
587 device
->master_fd
= master_fd
;
589 result
= anv_init_wsi(device
);
590 if (result
!= VK_SUCCESS
) {
591 ralloc_free(device
->compiler
);
592 anv_physical_device_free_disk_cache(device
);
596 anv_physical_device_get_supported_extensions(device
,
597 &device
->supported_extensions
);
600 device
->local_fd
= fd
;
612 anv_physical_device_finish(struct anv_physical_device
*device
)
614 anv_finish_wsi(device
);
615 anv_physical_device_free_disk_cache(device
);
616 ralloc_free(device
->compiler
);
617 close(device
->local_fd
);
618 if (device
->master_fd
>= 0)
619 close(device
->master_fd
);
623 default_alloc_func(void *pUserData
, size_t size
, size_t align
,
624 VkSystemAllocationScope allocationScope
)
630 default_realloc_func(void *pUserData
, void *pOriginal
, size_t size
,
631 size_t align
, VkSystemAllocationScope allocationScope
)
633 return realloc(pOriginal
, size
);
637 default_free_func(void *pUserData
, void *pMemory
)
642 static const VkAllocationCallbacks default_alloc
= {
644 .pfnAllocation
= default_alloc_func
,
645 .pfnReallocation
= default_realloc_func
,
646 .pfnFree
= default_free_func
,
649 VkResult
anv_EnumerateInstanceExtensionProperties(
650 const char* pLayerName
,
651 uint32_t* pPropertyCount
,
652 VkExtensionProperties
* pProperties
)
654 VK_OUTARRAY_MAKE(out
, pProperties
, pPropertyCount
);
656 for (int i
= 0; i
< ANV_INSTANCE_EXTENSION_COUNT
; i
++) {
657 if (anv_instance_extensions_supported
.extensions
[i
]) {
658 vk_outarray_append(&out
, prop
) {
659 *prop
= anv_instance_extensions
[i
];
664 return vk_outarray_status(&out
);
667 VkResult
anv_CreateInstance(
668 const VkInstanceCreateInfo
* pCreateInfo
,
669 const VkAllocationCallbacks
* pAllocator
,
670 VkInstance
* pInstance
)
672 struct anv_instance
*instance
;
675 assert(pCreateInfo
->sType
== VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO
);
677 struct anv_instance_extension_table enabled_extensions
= {};
678 for (uint32_t i
= 0; i
< pCreateInfo
->enabledExtensionCount
; i
++) {
680 for (idx
= 0; idx
< ANV_INSTANCE_EXTENSION_COUNT
; idx
++) {
681 if (strcmp(pCreateInfo
->ppEnabledExtensionNames
[i
],
682 anv_instance_extensions
[idx
].extensionName
) == 0)
686 if (idx
>= ANV_INSTANCE_EXTENSION_COUNT
)
687 return vk_error(VK_ERROR_EXTENSION_NOT_PRESENT
);
689 if (!anv_instance_extensions_supported
.extensions
[idx
])
690 return vk_error(VK_ERROR_EXTENSION_NOT_PRESENT
);
692 enabled_extensions
.extensions
[idx
] = true;
695 instance
= vk_alloc2(&default_alloc
, pAllocator
, sizeof(*instance
), 8,
696 VK_SYSTEM_ALLOCATION_SCOPE_INSTANCE
);
698 return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY
);
700 instance
->_loader_data
.loaderMagic
= ICD_LOADER_MAGIC
;
703 instance
->alloc
= *pAllocator
;
705 instance
->alloc
= default_alloc
;
707 instance
->app_info
= (struct anv_app_info
) { .api_version
= 0 };
708 if (pCreateInfo
->pApplicationInfo
) {
709 const VkApplicationInfo
*app
= pCreateInfo
->pApplicationInfo
;
711 instance
->app_info
.app_name
=
712 vk_strdup(&instance
->alloc
, app
->pApplicationName
,
713 VK_SYSTEM_ALLOCATION_SCOPE_INSTANCE
);
714 instance
->app_info
.app_version
= app
->applicationVersion
;
716 instance
->app_info
.engine_name
=
717 vk_strdup(&instance
->alloc
, app
->pEngineName
,
718 VK_SYSTEM_ALLOCATION_SCOPE_INSTANCE
);
719 instance
->app_info
.engine_version
= app
->engineVersion
;
721 instance
->app_info
.api_version
= app
->apiVersion
;
724 if (instance
->app_info
.api_version
== 0)
725 instance
->app_info
.api_version
= VK_API_VERSION_1_0
;
727 instance
->enabled_extensions
= enabled_extensions
;
729 for (unsigned i
= 0; i
< ARRAY_SIZE(instance
->dispatch
.entrypoints
); i
++) {
730 /* Vulkan requires that entrypoints for extensions which have not been
731 * enabled must not be advertised.
733 if (!anv_instance_entrypoint_is_enabled(i
, instance
->app_info
.api_version
,
734 &instance
->enabled_extensions
)) {
735 instance
->dispatch
.entrypoints
[i
] = NULL
;
737 instance
->dispatch
.entrypoints
[i
] =
738 anv_instance_dispatch_table
.entrypoints
[i
];
742 for (unsigned i
= 0; i
< ARRAY_SIZE(instance
->device_dispatch
.entrypoints
); i
++) {
743 /* Vulkan requires that entrypoints for extensions which have not been
744 * enabled must not be advertised.
746 if (!anv_device_entrypoint_is_enabled(i
, instance
->app_info
.api_version
,
747 &instance
->enabled_extensions
, NULL
)) {
748 instance
->device_dispatch
.entrypoints
[i
] = NULL
;
750 instance
->device_dispatch
.entrypoints
[i
] =
751 anv_device_dispatch_table
.entrypoints
[i
];
755 instance
->physicalDeviceCount
= -1;
757 result
= vk_debug_report_instance_init(&instance
->debug_report_callbacks
);
758 if (result
!= VK_SUCCESS
) {
759 vk_free2(&default_alloc
, pAllocator
, instance
);
760 return vk_error(result
);
763 instance
->pipeline_cache_enabled
=
764 env_var_as_boolean("ANV_ENABLE_PIPELINE_CACHE", true);
767 glsl_type_singleton_init_or_ref();
769 VG(VALGRIND_CREATE_MEMPOOL(instance
, 0, false));
771 *pInstance
= anv_instance_to_handle(instance
);
776 void anv_DestroyInstance(
777 VkInstance _instance
,
778 const VkAllocationCallbacks
* pAllocator
)
780 ANV_FROM_HANDLE(anv_instance
, instance
, _instance
);
785 if (instance
->physicalDeviceCount
> 0) {
786 /* We support at most one physical device. */
787 assert(instance
->physicalDeviceCount
== 1);
788 anv_physical_device_finish(&instance
->physicalDevice
);
791 vk_free(&instance
->alloc
, (char *)instance
->app_info
.app_name
);
792 vk_free(&instance
->alloc
, (char *)instance
->app_info
.engine_name
);
794 VG(VALGRIND_DESTROY_MEMPOOL(instance
));
796 vk_debug_report_instance_destroy(&instance
->debug_report_callbacks
);
798 glsl_type_singleton_decref();
801 vk_free(&instance
->alloc
, instance
);
805 anv_enumerate_devices(struct anv_instance
*instance
)
807 /* TODO: Check for more devices ? */
808 drmDevicePtr devices
[8];
809 VkResult result
= VK_ERROR_INCOMPATIBLE_DRIVER
;
812 instance
->physicalDeviceCount
= 0;
814 max_devices
= drmGetDevices2(0, devices
, ARRAY_SIZE(devices
));
816 return VK_ERROR_INCOMPATIBLE_DRIVER
;
818 for (unsigned i
= 0; i
< (unsigned)max_devices
; i
++) {
819 if (devices
[i
]->available_nodes
& 1 << DRM_NODE_RENDER
&&
820 devices
[i
]->bustype
== DRM_BUS_PCI
&&
821 devices
[i
]->deviceinfo
.pci
->vendor_id
== 0x8086) {
823 result
= anv_physical_device_init(&instance
->physicalDevice
,
824 instance
, devices
[i
]);
825 if (result
!= VK_ERROR_INCOMPATIBLE_DRIVER
)
829 drmFreeDevices(devices
, max_devices
);
831 if (result
== VK_SUCCESS
)
832 instance
->physicalDeviceCount
= 1;
838 anv_instance_ensure_physical_device(struct anv_instance
*instance
)
840 if (instance
->physicalDeviceCount
< 0) {
841 VkResult result
= anv_enumerate_devices(instance
);
842 if (result
!= VK_SUCCESS
&&
843 result
!= VK_ERROR_INCOMPATIBLE_DRIVER
)
850 VkResult
anv_EnumeratePhysicalDevices(
851 VkInstance _instance
,
852 uint32_t* pPhysicalDeviceCount
,
853 VkPhysicalDevice
* pPhysicalDevices
)
855 ANV_FROM_HANDLE(anv_instance
, instance
, _instance
);
856 VK_OUTARRAY_MAKE(out
, pPhysicalDevices
, pPhysicalDeviceCount
);
858 VkResult result
= anv_instance_ensure_physical_device(instance
);
859 if (result
!= VK_SUCCESS
)
862 if (instance
->physicalDeviceCount
== 0)
865 assert(instance
->physicalDeviceCount
== 1);
866 vk_outarray_append(&out
, i
) {
867 *i
= anv_physical_device_to_handle(&instance
->physicalDevice
);
870 return vk_outarray_status(&out
);
873 VkResult
anv_EnumeratePhysicalDeviceGroups(
874 VkInstance _instance
,
875 uint32_t* pPhysicalDeviceGroupCount
,
876 VkPhysicalDeviceGroupProperties
* pPhysicalDeviceGroupProperties
)
878 ANV_FROM_HANDLE(anv_instance
, instance
, _instance
);
879 VK_OUTARRAY_MAKE(out
, pPhysicalDeviceGroupProperties
,
880 pPhysicalDeviceGroupCount
);
882 VkResult result
= anv_instance_ensure_physical_device(instance
);
883 if (result
!= VK_SUCCESS
)
886 if (instance
->physicalDeviceCount
== 0)
889 assert(instance
->physicalDeviceCount
== 1);
891 vk_outarray_append(&out
, p
) {
892 p
->physicalDeviceCount
= 1;
893 memset(p
->physicalDevices
, 0, sizeof(p
->physicalDevices
));
894 p
->physicalDevices
[0] =
895 anv_physical_device_to_handle(&instance
->physicalDevice
);
896 p
->subsetAllocation
= false;
898 vk_foreach_struct(ext
, p
->pNext
)
899 anv_debug_ignored_stype(ext
->sType
);
902 return vk_outarray_status(&out
);
905 void anv_GetPhysicalDeviceFeatures(
906 VkPhysicalDevice physicalDevice
,
907 VkPhysicalDeviceFeatures
* pFeatures
)
909 ANV_FROM_HANDLE(anv_physical_device
, pdevice
, physicalDevice
);
911 *pFeatures
= (VkPhysicalDeviceFeatures
) {
912 .robustBufferAccess
= true,
913 .fullDrawIndexUint32
= true,
914 .imageCubeArray
= true,
915 .independentBlend
= true,
916 .geometryShader
= true,
917 .tessellationShader
= true,
918 .sampleRateShading
= true,
919 .dualSrcBlend
= true,
921 .multiDrawIndirect
= true,
922 .drawIndirectFirstInstance
= true,
924 .depthBiasClamp
= true,
925 .fillModeNonSolid
= true,
926 .depthBounds
= false,
930 .multiViewport
= true,
931 .samplerAnisotropy
= true,
932 .textureCompressionETC2
= pdevice
->info
.gen
>= 8 ||
933 pdevice
->info
.is_baytrail
,
934 .textureCompressionASTC_LDR
= pdevice
->info
.gen
>= 9, /* FINISHME CHV */
935 .textureCompressionBC
= true,
936 .occlusionQueryPrecise
= true,
937 .pipelineStatisticsQuery
= true,
938 .fragmentStoresAndAtomics
= true,
939 .shaderTessellationAndGeometryPointSize
= true,
940 .shaderImageGatherExtended
= true,
941 .shaderStorageImageExtendedFormats
= true,
942 .shaderStorageImageMultisample
= false,
943 .shaderStorageImageReadWithoutFormat
= false,
944 .shaderStorageImageWriteWithoutFormat
= true,
945 .shaderUniformBufferArrayDynamicIndexing
= true,
946 .shaderSampledImageArrayDynamicIndexing
= true,
947 .shaderStorageBufferArrayDynamicIndexing
= true,
948 .shaderStorageImageArrayDynamicIndexing
= true,
949 .shaderClipDistance
= true,
950 .shaderCullDistance
= true,
951 .shaderFloat64
= pdevice
->info
.gen
>= 8 &&
952 pdevice
->info
.has_64bit_types
,
953 .shaderInt64
= pdevice
->info
.gen
>= 8 &&
954 pdevice
->info
.has_64bit_types
,
955 .shaderInt16
= pdevice
->info
.gen
>= 8,
956 .shaderResourceMinLod
= pdevice
->info
.gen
>= 9,
957 .variableMultisampleRate
= true,
958 .inheritedQueries
= true,
961 /* We can't do image stores in vec4 shaders */
962 pFeatures
->vertexPipelineStoresAndAtomics
=
963 pdevice
->compiler
->scalar_stage
[MESA_SHADER_VERTEX
] &&
964 pdevice
->compiler
->scalar_stage
[MESA_SHADER_GEOMETRY
];
966 struct anv_app_info
*app_info
= &pdevice
->instance
->app_info
;
968 /* The new DOOM and Wolfenstein games require depthBounds without
969 * checking for it. They seem to run fine without it so just claim it's
970 * there and accept the consequences.
972 if (app_info
->engine_name
&& strcmp(app_info
->engine_name
, "idTech") == 0)
973 pFeatures
->depthBounds
= true;
976 void anv_GetPhysicalDeviceFeatures2(
977 VkPhysicalDevice physicalDevice
,
978 VkPhysicalDeviceFeatures2
* pFeatures
)
980 ANV_FROM_HANDLE(anv_physical_device
, pdevice
, physicalDevice
);
981 anv_GetPhysicalDeviceFeatures(physicalDevice
, &pFeatures
->features
);
983 vk_foreach_struct(ext
, pFeatures
->pNext
) {
984 switch (ext
->sType
) {
985 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_8BIT_STORAGE_FEATURES_KHR
: {
986 VkPhysicalDevice8BitStorageFeaturesKHR
*features
=
987 (VkPhysicalDevice8BitStorageFeaturesKHR
*)ext
;
988 features
->storageBuffer8BitAccess
= pdevice
->info
.gen
>= 8;
989 features
->uniformAndStorageBuffer8BitAccess
= pdevice
->info
.gen
>= 8;
990 features
->storagePushConstant8
= pdevice
->info
.gen
>= 8;
994 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_16BIT_STORAGE_FEATURES
: {
995 VkPhysicalDevice16BitStorageFeatures
*features
=
996 (VkPhysicalDevice16BitStorageFeatures
*)ext
;
997 features
->storageBuffer16BitAccess
= pdevice
->info
.gen
>= 8;
998 features
->uniformAndStorageBuffer16BitAccess
= pdevice
->info
.gen
>= 8;
999 features
->storagePushConstant16
= pdevice
->info
.gen
>= 8;
1000 features
->storageInputOutput16
= false;
1004 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_BUFFER_DEVICE_ADDRESS_FEATURES_EXT
: {
1005 VkPhysicalDeviceBufferDeviceAddressFeaturesEXT
*features
= (void *)ext
;
1006 features
->bufferDeviceAddress
= pdevice
->has_a64_buffer_access
;
1007 features
->bufferDeviceAddressCaptureReplay
= false;
1008 features
->bufferDeviceAddressMultiDevice
= false;
1012 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_COMPUTE_SHADER_DERIVATIVES_FEATURES_NV
: {
1013 VkPhysicalDeviceComputeShaderDerivativesFeaturesNV
*features
=
1014 (VkPhysicalDeviceComputeShaderDerivativesFeaturesNV
*)ext
;
1015 features
->computeDerivativeGroupQuads
= true;
1016 features
->computeDerivativeGroupLinear
= true;
1020 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_CONDITIONAL_RENDERING_FEATURES_EXT
: {
1021 VkPhysicalDeviceConditionalRenderingFeaturesEXT
*features
=
1022 (VkPhysicalDeviceConditionalRenderingFeaturesEXT
*)ext
;
1023 features
->conditionalRendering
= pdevice
->info
.gen
>= 8 ||
1024 pdevice
->info
.is_haswell
;
1025 features
->inheritedConditionalRendering
= pdevice
->info
.gen
>= 8 ||
1026 pdevice
->info
.is_haswell
;
1030 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DEPTH_CLIP_ENABLE_FEATURES_EXT
: {
1031 VkPhysicalDeviceDepthClipEnableFeaturesEXT
*features
=
1032 (VkPhysicalDeviceDepthClipEnableFeaturesEXT
*)ext
;
1033 features
->depthClipEnable
= true;
1037 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FLOAT16_INT8_FEATURES_KHR
: {
1038 VkPhysicalDeviceFloat16Int8FeaturesKHR
*features
= (void *)ext
;
1039 features
->shaderFloat16
= pdevice
->info
.gen
>= 8;
1040 features
->shaderInt8
= pdevice
->info
.gen
>= 8;
1044 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FRAGMENT_SHADER_INTERLOCK_FEATURES_EXT
: {
1045 VkPhysicalDeviceFragmentShaderInterlockFeaturesEXT
*features
=
1046 (VkPhysicalDeviceFragmentShaderInterlockFeaturesEXT
*)ext
;
1047 features
->fragmentShaderSampleInterlock
= pdevice
->info
.gen
>= 9;
1048 features
->fragmentShaderPixelInterlock
= pdevice
->info
.gen
>= 9;
1049 features
->fragmentShaderShadingRateInterlock
= false;
1053 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_HOST_QUERY_RESET_FEATURES_EXT
: {
1054 VkPhysicalDeviceHostQueryResetFeaturesEXT
*features
=
1055 (VkPhysicalDeviceHostQueryResetFeaturesEXT
*)ext
;
1056 features
->hostQueryReset
= true;
1060 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DESCRIPTOR_INDEXING_FEATURES_EXT
: {
1061 VkPhysicalDeviceDescriptorIndexingFeaturesEXT
*features
=
1062 (VkPhysicalDeviceDescriptorIndexingFeaturesEXT
*)ext
;
1063 features
->shaderInputAttachmentArrayDynamicIndexing
= false;
1064 features
->shaderUniformTexelBufferArrayDynamicIndexing
= true;
1065 features
->shaderStorageTexelBufferArrayDynamicIndexing
= true;
1066 features
->shaderUniformBufferArrayNonUniformIndexing
= false;
1067 features
->shaderSampledImageArrayNonUniformIndexing
= true;
1068 features
->shaderStorageBufferArrayNonUniformIndexing
= true;
1069 features
->shaderStorageImageArrayNonUniformIndexing
= true;
1070 features
->shaderInputAttachmentArrayNonUniformIndexing
= false;
1071 features
->shaderUniformTexelBufferArrayNonUniformIndexing
= true;
1072 features
->shaderStorageTexelBufferArrayNonUniformIndexing
= true;
1073 features
->descriptorBindingUniformBufferUpdateAfterBind
= false;
1074 features
->descriptorBindingSampledImageUpdateAfterBind
= true;
1075 features
->descriptorBindingStorageImageUpdateAfterBind
= true;
1076 features
->descriptorBindingStorageBufferUpdateAfterBind
= true;
1077 features
->descriptorBindingUniformTexelBufferUpdateAfterBind
= true;
1078 features
->descriptorBindingStorageTexelBufferUpdateAfterBind
= true;
1079 features
->descriptorBindingUpdateUnusedWhilePending
= true;
1080 features
->descriptorBindingPartiallyBound
= true;
1081 features
->descriptorBindingVariableDescriptorCount
= false;
1082 features
->runtimeDescriptorArray
= true;
1086 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_INDEX_TYPE_UINT8_FEATURES_EXT
: {
1087 VkPhysicalDeviceIndexTypeUint8FeaturesEXT
*features
=
1088 (VkPhysicalDeviceIndexTypeUint8FeaturesEXT
*)ext
;
1089 features
->indexTypeUint8
= true;
1093 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_INLINE_UNIFORM_BLOCK_FEATURES_EXT
: {
1094 VkPhysicalDeviceInlineUniformBlockFeaturesEXT
*features
=
1095 (VkPhysicalDeviceInlineUniformBlockFeaturesEXT
*)ext
;
1096 features
->inlineUniformBlock
= true;
1097 features
->descriptorBindingInlineUniformBlockUpdateAfterBind
= true;
1101 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_LINE_RASTERIZATION_FEATURES_EXT
: {
1102 VkPhysicalDeviceLineRasterizationFeaturesEXT
*features
=
1103 (VkPhysicalDeviceLineRasterizationFeaturesEXT
*)ext
;
1104 features
->rectangularLines
= true;
1105 features
->bresenhamLines
= true;
1106 features
->smoothLines
= true;
1107 features
->stippledRectangularLines
= false;
1108 features
->stippledBresenhamLines
= true;
1109 features
->stippledSmoothLines
= false;
1113 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_MULTIVIEW_FEATURES
: {
1114 VkPhysicalDeviceMultiviewFeatures
*features
=
1115 (VkPhysicalDeviceMultiviewFeatures
*)ext
;
1116 features
->multiview
= true;
1117 features
->multiviewGeometryShader
= true;
1118 features
->multiviewTessellationShader
= true;
1122 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_IMAGELESS_FRAMEBUFFER_FEATURES_KHR
: {
1123 VkPhysicalDeviceImagelessFramebufferFeaturesKHR
*features
=
1124 (VkPhysicalDeviceImagelessFramebufferFeaturesKHR
*)ext
;
1125 features
->imagelessFramebuffer
= true;
1129 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROTECTED_MEMORY_FEATURES
: {
1130 VkPhysicalDeviceProtectedMemoryFeatures
*features
= (void *)ext
;
1131 features
->protectedMemory
= false;
1135 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SAMPLER_YCBCR_CONVERSION_FEATURES
: {
1136 VkPhysicalDeviceSamplerYcbcrConversionFeatures
*features
=
1137 (VkPhysicalDeviceSamplerYcbcrConversionFeatures
*) ext
;
1138 features
->samplerYcbcrConversion
= true;
1142 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SCALAR_BLOCK_LAYOUT_FEATURES_EXT
: {
1143 VkPhysicalDeviceScalarBlockLayoutFeaturesEXT
*features
=
1144 (VkPhysicalDeviceScalarBlockLayoutFeaturesEXT
*)ext
;
1145 features
->scalarBlockLayout
= true;
1149 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SHADER_ATOMIC_INT64_FEATURES_KHR
: {
1150 VkPhysicalDeviceShaderAtomicInt64FeaturesKHR
*features
= (void *)ext
;
1151 features
->shaderBufferInt64Atomics
=
1152 pdevice
->info
.gen
>= 9 && pdevice
->use_softpin
;
1153 features
->shaderSharedInt64Atomics
= VK_FALSE
;
1157 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SHADER_DEMOTE_TO_HELPER_INVOCATION_FEATURES_EXT
: {
1158 VkPhysicalDeviceShaderDemoteToHelperInvocationFeaturesEXT
*features
= (void *)ext
;
1159 features
->shaderDemoteToHelperInvocation
= true;
1163 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SHADER_DRAW_PARAMETERS_FEATURES
: {
1164 VkPhysicalDeviceShaderDrawParametersFeatures
*features
= (void *)ext
;
1165 features
->shaderDrawParameters
= true;
1169 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SUBGROUP_SIZE_CONTROL_FEATURES_EXT
: {
1170 VkPhysicalDeviceSubgroupSizeControlFeaturesEXT
*features
=
1171 (VkPhysicalDeviceSubgroupSizeControlFeaturesEXT
*)ext
;
1172 features
->subgroupSizeControl
= true;
1173 features
->computeFullSubgroups
= true;
1177 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TEXEL_BUFFER_ALIGNMENT_FEATURES_EXT
: {
1178 VkPhysicalDeviceTexelBufferAlignmentFeaturesEXT
*features
=
1179 (VkPhysicalDeviceTexelBufferAlignmentFeaturesEXT
*)ext
;
1180 features
->texelBufferAlignment
= true;
1184 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VARIABLE_POINTERS_FEATURES
: {
1185 VkPhysicalDeviceVariablePointersFeatures
*features
= (void *)ext
;
1186 features
->variablePointersStorageBuffer
= true;
1187 features
->variablePointers
= true;
1191 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TRANSFORM_FEEDBACK_FEATURES_EXT
: {
1192 VkPhysicalDeviceTransformFeedbackFeaturesEXT
*features
=
1193 (VkPhysicalDeviceTransformFeedbackFeaturesEXT
*)ext
;
1194 features
->transformFeedback
= true;
1195 features
->geometryStreams
= true;
1199 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_UNIFORM_BUFFER_STANDARD_LAYOUT_FEATURES_KHR
: {
1200 VkPhysicalDeviceUniformBufferStandardLayoutFeaturesKHR
*features
=
1201 (VkPhysicalDeviceUniformBufferStandardLayoutFeaturesKHR
*)ext
;
1202 features
->uniformBufferStandardLayout
= true;
1206 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VERTEX_ATTRIBUTE_DIVISOR_FEATURES_EXT
: {
1207 VkPhysicalDeviceVertexAttributeDivisorFeaturesEXT
*features
=
1208 (VkPhysicalDeviceVertexAttributeDivisorFeaturesEXT
*)ext
;
1209 features
->vertexAttributeInstanceRateDivisor
= true;
1210 features
->vertexAttributeInstanceRateZeroDivisor
= true;
1214 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_YCBCR_IMAGE_ARRAYS_FEATURES_EXT
: {
1215 VkPhysicalDeviceYcbcrImageArraysFeaturesEXT
*features
=
1216 (VkPhysicalDeviceYcbcrImageArraysFeaturesEXT
*)ext
;
1217 features
->ycbcrImageArrays
= true;
1222 anv_debug_ignored_stype(ext
->sType
);
1228 #define MAX_PER_STAGE_DESCRIPTOR_UNIFORM_BUFFERS 64
1230 #define MAX_PER_STAGE_DESCRIPTOR_INPUT_ATTACHMENTS 64
1231 #define MAX_DESCRIPTOR_SET_INPUT_ATTACHMENTS 256
1233 void anv_GetPhysicalDeviceProperties(
1234 VkPhysicalDevice physicalDevice
,
1235 VkPhysicalDeviceProperties
* pProperties
)
1237 ANV_FROM_HANDLE(anv_physical_device
, pdevice
, physicalDevice
);
1238 const struct gen_device_info
*devinfo
= &pdevice
->info
;
1240 /* See assertions made when programming the buffer surface state. */
1241 const uint32_t max_raw_buffer_sz
= devinfo
->gen
>= 7 ?
1242 (1ul << 30) : (1ul << 27);
1244 const uint32_t max_ssbos
= pdevice
->has_a64_buffer_access
? UINT16_MAX
: 64;
1245 const uint32_t max_textures
=
1246 pdevice
->has_bindless_images
? UINT16_MAX
: 128;
1247 const uint32_t max_samplers
=
1248 pdevice
->has_bindless_samplers
? UINT16_MAX
:
1249 (devinfo
->gen
>= 8 || devinfo
->is_haswell
) ? 128 : 16;
1250 const uint32_t max_images
=
1251 pdevice
->has_bindless_images
? UINT16_MAX
: MAX_IMAGES
;
1253 /* The moment we have anything bindless, claim a high per-stage limit */
1254 const uint32_t max_per_stage
=
1255 pdevice
->has_a64_buffer_access
? UINT32_MAX
:
1256 MAX_BINDING_TABLE_SIZE
- MAX_RTS
;
1258 VkSampleCountFlags sample_counts
=
1259 isl_device_get_sample_counts(&pdevice
->isl_dev
);
1262 VkPhysicalDeviceLimits limits
= {
1263 .maxImageDimension1D
= (1 << 14),
1264 .maxImageDimension2D
= (1 << 14),
1265 .maxImageDimension3D
= (1 << 11),
1266 .maxImageDimensionCube
= (1 << 14),
1267 .maxImageArrayLayers
= (1 << 11),
1268 .maxTexelBufferElements
= 128 * 1024 * 1024,
1269 .maxUniformBufferRange
= (1ul << 27),
1270 .maxStorageBufferRange
= max_raw_buffer_sz
,
1271 .maxPushConstantsSize
= MAX_PUSH_CONSTANTS_SIZE
,
1272 .maxMemoryAllocationCount
= UINT32_MAX
,
1273 .maxSamplerAllocationCount
= 64 * 1024,
1274 .bufferImageGranularity
= 64, /* A cache line */
1275 .sparseAddressSpaceSize
= 0,
1276 .maxBoundDescriptorSets
= MAX_SETS
,
1277 .maxPerStageDescriptorSamplers
= max_samplers
,
1278 .maxPerStageDescriptorUniformBuffers
= MAX_PER_STAGE_DESCRIPTOR_UNIFORM_BUFFERS
,
1279 .maxPerStageDescriptorStorageBuffers
= max_ssbos
,
1280 .maxPerStageDescriptorSampledImages
= max_textures
,
1281 .maxPerStageDescriptorStorageImages
= max_images
,
1282 .maxPerStageDescriptorInputAttachments
= MAX_PER_STAGE_DESCRIPTOR_INPUT_ATTACHMENTS
,
1283 .maxPerStageResources
= max_per_stage
,
1284 .maxDescriptorSetSamplers
= 6 * max_samplers
, /* number of stages * maxPerStageDescriptorSamplers */
1285 .maxDescriptorSetUniformBuffers
= 6 * MAX_PER_STAGE_DESCRIPTOR_UNIFORM_BUFFERS
, /* number of stages * maxPerStageDescriptorUniformBuffers */
1286 .maxDescriptorSetUniformBuffersDynamic
= MAX_DYNAMIC_BUFFERS
/ 2,
1287 .maxDescriptorSetStorageBuffers
= 6 * max_ssbos
, /* number of stages * maxPerStageDescriptorStorageBuffers */
1288 .maxDescriptorSetStorageBuffersDynamic
= MAX_DYNAMIC_BUFFERS
/ 2,
1289 .maxDescriptorSetSampledImages
= 6 * max_textures
, /* number of stages * maxPerStageDescriptorSampledImages */
1290 .maxDescriptorSetStorageImages
= 6 * max_images
, /* number of stages * maxPerStageDescriptorStorageImages */
1291 .maxDescriptorSetInputAttachments
= MAX_DESCRIPTOR_SET_INPUT_ATTACHMENTS
,
1292 .maxVertexInputAttributes
= MAX_VBS
,
1293 .maxVertexInputBindings
= MAX_VBS
,
1294 .maxVertexInputAttributeOffset
= 2047,
1295 .maxVertexInputBindingStride
= 2048,
1296 .maxVertexOutputComponents
= 128,
1297 .maxTessellationGenerationLevel
= 64,
1298 .maxTessellationPatchSize
= 32,
1299 .maxTessellationControlPerVertexInputComponents
= 128,
1300 .maxTessellationControlPerVertexOutputComponents
= 128,
1301 .maxTessellationControlPerPatchOutputComponents
= 128,
1302 .maxTessellationControlTotalOutputComponents
= 2048,
1303 .maxTessellationEvaluationInputComponents
= 128,
1304 .maxTessellationEvaluationOutputComponents
= 128,
1305 .maxGeometryShaderInvocations
= 32,
1306 .maxGeometryInputComponents
= 64,
1307 .maxGeometryOutputComponents
= 128,
1308 .maxGeometryOutputVertices
= 256,
1309 .maxGeometryTotalOutputComponents
= 1024,
1310 .maxFragmentInputComponents
= 116, /* 128 components - (PSIZ, CLIP_DIST0, CLIP_DIST1) */
1311 .maxFragmentOutputAttachments
= 8,
1312 .maxFragmentDualSrcAttachments
= 1,
1313 .maxFragmentCombinedOutputResources
= 8,
1314 .maxComputeSharedMemorySize
= 64 * 1024,
1315 .maxComputeWorkGroupCount
= { 65535, 65535, 65535 },
1316 .maxComputeWorkGroupInvocations
= 32 * devinfo
->max_cs_threads
,
1317 .maxComputeWorkGroupSize
= {
1318 16 * devinfo
->max_cs_threads
,
1319 16 * devinfo
->max_cs_threads
,
1320 16 * devinfo
->max_cs_threads
,
1322 .subPixelPrecisionBits
= 8,
1323 .subTexelPrecisionBits
= 8,
1324 .mipmapPrecisionBits
= 8,
1325 .maxDrawIndexedIndexValue
= UINT32_MAX
,
1326 .maxDrawIndirectCount
= UINT32_MAX
,
1327 .maxSamplerLodBias
= 16,
1328 .maxSamplerAnisotropy
= 16,
1329 .maxViewports
= MAX_VIEWPORTS
,
1330 .maxViewportDimensions
= { (1 << 14), (1 << 14) },
1331 .viewportBoundsRange
= { INT16_MIN
, INT16_MAX
},
1332 .viewportSubPixelBits
= 13, /* We take a float? */
1333 .minMemoryMapAlignment
= 4096, /* A page */
1334 /* The dataport requires texel alignment so we need to assume a worst
1335 * case of R32G32B32A32 which is 16 bytes.
1337 .minTexelBufferOffsetAlignment
= 16,
1338 /* We need 16 for UBO block reads to work and 32 for push UBOs */
1339 .minUniformBufferOffsetAlignment
= 32,
1340 .minStorageBufferOffsetAlignment
= 4,
1341 .minTexelOffset
= -8,
1342 .maxTexelOffset
= 7,
1343 .minTexelGatherOffset
= -32,
1344 .maxTexelGatherOffset
= 31,
1345 .minInterpolationOffset
= -0.5,
1346 .maxInterpolationOffset
= 0.4375,
1347 .subPixelInterpolationOffsetBits
= 4,
1348 .maxFramebufferWidth
= (1 << 14),
1349 .maxFramebufferHeight
= (1 << 14),
1350 .maxFramebufferLayers
= (1 << 11),
1351 .framebufferColorSampleCounts
= sample_counts
,
1352 .framebufferDepthSampleCounts
= sample_counts
,
1353 .framebufferStencilSampleCounts
= sample_counts
,
1354 .framebufferNoAttachmentsSampleCounts
= sample_counts
,
1355 .maxColorAttachments
= MAX_RTS
,
1356 .sampledImageColorSampleCounts
= sample_counts
,
1357 .sampledImageIntegerSampleCounts
= VK_SAMPLE_COUNT_1_BIT
,
1358 .sampledImageDepthSampleCounts
= sample_counts
,
1359 .sampledImageStencilSampleCounts
= sample_counts
,
1360 .storageImageSampleCounts
= VK_SAMPLE_COUNT_1_BIT
,
1361 .maxSampleMaskWords
= 1,
1362 .timestampComputeAndGraphics
= true,
1363 .timestampPeriod
= 1000000000.0 / devinfo
->timestamp_frequency
,
1364 .maxClipDistances
= 8,
1365 .maxCullDistances
= 8,
1366 .maxCombinedClipAndCullDistances
= 8,
1367 .discreteQueuePriorities
= 2,
1368 .pointSizeRange
= { 0.125, 255.875 },
1371 (devinfo
->gen
>= 9 || devinfo
->is_cherryview
) ?
1372 2047.9921875 : 7.9921875,
1374 .pointSizeGranularity
= (1.0 / 8.0),
1375 .lineWidthGranularity
= (1.0 / 128.0),
1376 .strictLines
= false,
1377 .standardSampleLocations
= true,
1378 .optimalBufferCopyOffsetAlignment
= 128,
1379 .optimalBufferCopyRowPitchAlignment
= 128,
1380 .nonCoherentAtomSize
= 64,
1383 *pProperties
= (VkPhysicalDeviceProperties
) {
1384 .apiVersion
= anv_physical_device_api_version(pdevice
),
1385 .driverVersion
= vk_get_driver_version(),
1387 .deviceID
= pdevice
->chipset_id
,
1388 .deviceType
= VK_PHYSICAL_DEVICE_TYPE_INTEGRATED_GPU
,
1390 .sparseProperties
= {0}, /* Broadwell doesn't do sparse. */
1393 snprintf(pProperties
->deviceName
, sizeof(pProperties
->deviceName
),
1394 "%s", pdevice
->name
);
1395 memcpy(pProperties
->pipelineCacheUUID
,
1396 pdevice
->pipeline_cache_uuid
, VK_UUID_SIZE
);
1399 void anv_GetPhysicalDeviceProperties2(
1400 VkPhysicalDevice physicalDevice
,
1401 VkPhysicalDeviceProperties2
* pProperties
)
1403 ANV_FROM_HANDLE(anv_physical_device
, pdevice
, physicalDevice
);
1405 anv_GetPhysicalDeviceProperties(physicalDevice
, &pProperties
->properties
);
1407 vk_foreach_struct(ext
, pProperties
->pNext
) {
1408 switch (ext
->sType
) {
1409 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DEPTH_STENCIL_RESOLVE_PROPERTIES_KHR
: {
1410 VkPhysicalDeviceDepthStencilResolvePropertiesKHR
*props
=
1411 (VkPhysicalDeviceDepthStencilResolvePropertiesKHR
*)ext
;
1413 /* We support all of the depth resolve modes */
1414 props
->supportedDepthResolveModes
=
1415 VK_RESOLVE_MODE_SAMPLE_ZERO_BIT_KHR
|
1416 VK_RESOLVE_MODE_AVERAGE_BIT_KHR
|
1417 VK_RESOLVE_MODE_MIN_BIT_KHR
|
1418 VK_RESOLVE_MODE_MAX_BIT_KHR
;
1420 /* Average doesn't make sense for stencil so we don't support that */
1421 props
->supportedStencilResolveModes
=
1422 VK_RESOLVE_MODE_SAMPLE_ZERO_BIT_KHR
;
1423 if (pdevice
->info
.gen
>= 8) {
1424 /* The advanced stencil resolve modes currently require stencil
1425 * sampling be supported by the hardware.
1427 props
->supportedStencilResolveModes
|=
1428 VK_RESOLVE_MODE_MIN_BIT_KHR
|
1429 VK_RESOLVE_MODE_MAX_BIT_KHR
;
1432 props
->independentResolveNone
= true;
1433 props
->independentResolve
= true;
1437 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DESCRIPTOR_INDEXING_PROPERTIES_EXT
: {
1438 VkPhysicalDeviceDescriptorIndexingPropertiesEXT
*props
=
1439 (VkPhysicalDeviceDescriptorIndexingPropertiesEXT
*)ext
;
1441 /* It's a bit hard to exactly map our implementation to the limits
1442 * described here. The bindless surface handle in the extended
1443 * message descriptors is 20 bits and it's an index into the table of
1444 * RENDER_SURFACE_STATE structs that starts at bindless surface base
1445 * address. Given that most things consume two surface states per
1446 * view (general/sampled for textures and write-only/read-write for
1447 * images), we claim 2^19 things.
1449 * For SSBOs, we just use A64 messages so there is no real limit
1450 * there beyond the limit on the total size of a descriptor set.
1452 const unsigned max_bindless_views
= 1 << 19;
1454 props
->maxUpdateAfterBindDescriptorsInAllPools
= max_bindless_views
;
1455 props
->shaderUniformBufferArrayNonUniformIndexingNative
= false;
1456 props
->shaderSampledImageArrayNonUniformIndexingNative
= false;
1457 props
->shaderStorageBufferArrayNonUniformIndexingNative
= true;
1458 props
->shaderStorageImageArrayNonUniformIndexingNative
= false;
1459 props
->shaderInputAttachmentArrayNonUniformIndexingNative
= false;
1460 props
->robustBufferAccessUpdateAfterBind
= true;
1461 props
->quadDivergentImplicitLod
= false;
1462 props
->maxPerStageDescriptorUpdateAfterBindSamplers
= max_bindless_views
;
1463 props
->maxPerStageDescriptorUpdateAfterBindUniformBuffers
= MAX_PER_STAGE_DESCRIPTOR_UNIFORM_BUFFERS
;
1464 props
->maxPerStageDescriptorUpdateAfterBindStorageBuffers
= UINT32_MAX
;
1465 props
->maxPerStageDescriptorUpdateAfterBindSampledImages
= max_bindless_views
;
1466 props
->maxPerStageDescriptorUpdateAfterBindStorageImages
= max_bindless_views
;
1467 props
->maxPerStageDescriptorUpdateAfterBindInputAttachments
= MAX_PER_STAGE_DESCRIPTOR_INPUT_ATTACHMENTS
;
1468 props
->maxPerStageUpdateAfterBindResources
= UINT32_MAX
;
1469 props
->maxDescriptorSetUpdateAfterBindSamplers
= max_bindless_views
;
1470 props
->maxDescriptorSetUpdateAfterBindUniformBuffers
= 6 * MAX_PER_STAGE_DESCRIPTOR_UNIFORM_BUFFERS
;
1471 props
->maxDescriptorSetUpdateAfterBindUniformBuffersDynamic
= MAX_DYNAMIC_BUFFERS
/ 2;
1472 props
->maxDescriptorSetUpdateAfterBindStorageBuffers
= UINT32_MAX
;
1473 props
->maxDescriptorSetUpdateAfterBindStorageBuffersDynamic
= MAX_DYNAMIC_BUFFERS
/ 2;
1474 props
->maxDescriptorSetUpdateAfterBindSampledImages
= max_bindless_views
;
1475 props
->maxDescriptorSetUpdateAfterBindStorageImages
= max_bindless_views
;
1476 props
->maxDescriptorSetUpdateAfterBindInputAttachments
= MAX_DESCRIPTOR_SET_INPUT_ATTACHMENTS
;
1480 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DRIVER_PROPERTIES_KHR
: {
1481 VkPhysicalDeviceDriverPropertiesKHR
*driver_props
=
1482 (VkPhysicalDeviceDriverPropertiesKHR
*) ext
;
1484 driver_props
->driverID
= VK_DRIVER_ID_INTEL_OPEN_SOURCE_MESA_KHR
;
1485 snprintf(driver_props
->driverName
, VK_MAX_DRIVER_NAME_SIZE_KHR
,
1486 "Intel open-source Mesa driver");
1488 snprintf(driver_props
->driverInfo
, VK_MAX_DRIVER_INFO_SIZE_KHR
,
1489 "Mesa " PACKAGE_VERSION MESA_GIT_SHA1
);
1491 driver_props
->conformanceVersion
= (VkConformanceVersionKHR
) {
1500 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_EXTERNAL_MEMORY_HOST_PROPERTIES_EXT
: {
1501 VkPhysicalDeviceExternalMemoryHostPropertiesEXT
*props
=
1502 (VkPhysicalDeviceExternalMemoryHostPropertiesEXT
*) ext
;
1503 /* Userptr needs page aligned memory. */
1504 props
->minImportedHostPointerAlignment
= 4096;
1508 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_ID_PROPERTIES
: {
1509 VkPhysicalDeviceIDProperties
*id_props
=
1510 (VkPhysicalDeviceIDProperties
*)ext
;
1511 memcpy(id_props
->deviceUUID
, pdevice
->device_uuid
, VK_UUID_SIZE
);
1512 memcpy(id_props
->driverUUID
, pdevice
->driver_uuid
, VK_UUID_SIZE
);
1513 /* The LUID is for Windows. */
1514 id_props
->deviceLUIDValid
= false;
1518 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_INLINE_UNIFORM_BLOCK_PROPERTIES_EXT
: {
1519 VkPhysicalDeviceInlineUniformBlockPropertiesEXT
*props
=
1520 (VkPhysicalDeviceInlineUniformBlockPropertiesEXT
*)ext
;
1521 props
->maxInlineUniformBlockSize
= MAX_INLINE_UNIFORM_BLOCK_SIZE
;
1522 props
->maxPerStageDescriptorInlineUniformBlocks
=
1523 MAX_INLINE_UNIFORM_BLOCK_DESCRIPTORS
;
1524 props
->maxPerStageDescriptorUpdateAfterBindInlineUniformBlocks
=
1525 MAX_INLINE_UNIFORM_BLOCK_DESCRIPTORS
;
1526 props
->maxDescriptorSetInlineUniformBlocks
=
1527 MAX_INLINE_UNIFORM_BLOCK_DESCRIPTORS
;
1528 props
->maxDescriptorSetUpdateAfterBindInlineUniformBlocks
=
1529 MAX_INLINE_UNIFORM_BLOCK_DESCRIPTORS
;
1533 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_LINE_RASTERIZATION_PROPERTIES_EXT
: {
1534 VkPhysicalDeviceLineRasterizationPropertiesEXT
*props
=
1535 (VkPhysicalDeviceLineRasterizationPropertiesEXT
*)ext
;
1536 /* In the Skylake PRM Vol. 7, subsection titled "GIQ (Diamond)
1537 * Sampling Rules - Legacy Mode", it says the following:
1539 * "Note that the device divides a pixel into a 16x16 array of
1540 * subpixels, referenced by their upper left corners."
1542 * This is the only known reference in the PRMs to the subpixel
1543 * precision of line rasterization and a "16x16 array of subpixels"
1544 * implies 4 subpixel precision bits. Empirical testing has shown
1545 * that 4 subpixel precision bits applies to all line rasterization
1548 props
->lineSubPixelPrecisionBits
= 4;
1552 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_MAINTENANCE_3_PROPERTIES
: {
1553 VkPhysicalDeviceMaintenance3Properties
*props
=
1554 (VkPhysicalDeviceMaintenance3Properties
*)ext
;
1555 /* This value doesn't matter for us today as our per-stage
1556 * descriptors are the real limit.
1558 props
->maxPerSetDescriptors
= 1024;
1559 props
->maxMemoryAllocationSize
= MAX_MEMORY_ALLOCATION_SIZE
;
1563 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_MULTIVIEW_PROPERTIES
: {
1564 VkPhysicalDeviceMultiviewProperties
*properties
=
1565 (VkPhysicalDeviceMultiviewProperties
*)ext
;
1566 properties
->maxMultiviewViewCount
= 16;
1567 properties
->maxMultiviewInstanceIndex
= UINT32_MAX
/ 16;
1571 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PCI_BUS_INFO_PROPERTIES_EXT
: {
1572 VkPhysicalDevicePCIBusInfoPropertiesEXT
*properties
=
1573 (VkPhysicalDevicePCIBusInfoPropertiesEXT
*)ext
;
1574 properties
->pciDomain
= pdevice
->pci_info
.domain
;
1575 properties
->pciBus
= pdevice
->pci_info
.bus
;
1576 properties
->pciDevice
= pdevice
->pci_info
.device
;
1577 properties
->pciFunction
= pdevice
->pci_info
.function
;
1581 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_POINT_CLIPPING_PROPERTIES
: {
1582 VkPhysicalDevicePointClippingProperties
*properties
=
1583 (VkPhysicalDevicePointClippingProperties
*) ext
;
1584 properties
->pointClippingBehavior
= VK_POINT_CLIPPING_BEHAVIOR_USER_CLIP_PLANES_ONLY
;
1588 #pragma GCC diagnostic push
1589 #pragma GCC diagnostic ignored "-Wswitch"
1590 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PRESENTATION_PROPERTIES_ANDROID
: {
1591 VkPhysicalDevicePresentationPropertiesANDROID
*props
=
1592 (VkPhysicalDevicePresentationPropertiesANDROID
*)ext
;
1593 props
->sharedImage
= VK_FALSE
;
1596 #pragma GCC diagnostic pop
1598 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROTECTED_MEMORY_PROPERTIES
: {
1599 VkPhysicalDeviceProtectedMemoryProperties
*props
=
1600 (VkPhysicalDeviceProtectedMemoryProperties
*)ext
;
1601 props
->protectedNoFault
= false;
1605 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PUSH_DESCRIPTOR_PROPERTIES_KHR
: {
1606 VkPhysicalDevicePushDescriptorPropertiesKHR
*properties
=
1607 (VkPhysicalDevicePushDescriptorPropertiesKHR
*) ext
;
1609 properties
->maxPushDescriptors
= MAX_PUSH_DESCRIPTORS
;
1613 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SAMPLER_FILTER_MINMAX_PROPERTIES_EXT
: {
1614 VkPhysicalDeviceSamplerFilterMinmaxPropertiesEXT
*properties
=
1615 (VkPhysicalDeviceSamplerFilterMinmaxPropertiesEXT
*)ext
;
1616 properties
->filterMinmaxImageComponentMapping
= pdevice
->info
.gen
>= 9;
1617 properties
->filterMinmaxSingleComponentFormats
= true;
1621 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SUBGROUP_PROPERTIES
: {
1622 VkPhysicalDeviceSubgroupProperties
*properties
= (void *)ext
;
1624 properties
->subgroupSize
= BRW_SUBGROUP_SIZE
;
1626 VkShaderStageFlags scalar_stages
= 0;
1627 for (unsigned stage
= 0; stage
< MESA_SHADER_STAGES
; stage
++) {
1628 if (pdevice
->compiler
->scalar_stage
[stage
])
1629 scalar_stages
|= mesa_to_vk_shader_stage(stage
);
1631 properties
->supportedStages
= scalar_stages
;
1633 properties
->supportedOperations
= VK_SUBGROUP_FEATURE_BASIC_BIT
|
1634 VK_SUBGROUP_FEATURE_VOTE_BIT
|
1635 VK_SUBGROUP_FEATURE_BALLOT_BIT
|
1636 VK_SUBGROUP_FEATURE_SHUFFLE_BIT
|
1637 VK_SUBGROUP_FEATURE_SHUFFLE_RELATIVE_BIT
|
1638 VK_SUBGROUP_FEATURE_QUAD_BIT
;
1639 if (pdevice
->info
.gen
>= 8) {
1640 /* TODO: There's no technical reason why these can't be made to
1641 * work on gen7 but they don't at the moment so it's best to leave
1642 * the feature disabled than enabled and broken.
1644 properties
->supportedOperations
|=
1645 VK_SUBGROUP_FEATURE_ARITHMETIC_BIT
|
1646 VK_SUBGROUP_FEATURE_CLUSTERED_BIT
;
1648 properties
->quadOperationsInAllStages
= pdevice
->info
.gen
>= 8;
1652 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SUBGROUP_SIZE_CONTROL_PROPERTIES_EXT
: {
1653 VkPhysicalDeviceSubgroupSizeControlPropertiesEXT
*props
=
1654 (VkPhysicalDeviceSubgroupSizeControlPropertiesEXT
*)ext
;
1655 STATIC_ASSERT(8 <= BRW_SUBGROUP_SIZE
&& BRW_SUBGROUP_SIZE
<= 32);
1656 props
->minSubgroupSize
= 8;
1657 props
->maxSubgroupSize
= 32;
1658 props
->maxComputeWorkgroupSubgroups
= pdevice
->info
.max_cs_threads
;
1659 props
->requiredSubgroupSizeStages
= VK_SHADER_STAGE_COMPUTE_BIT
;
1663 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TEXEL_BUFFER_ALIGNMENT_PROPERTIES_EXT
: {
1664 VkPhysicalDeviceTexelBufferAlignmentPropertiesEXT
*props
=
1665 (VkPhysicalDeviceTexelBufferAlignmentPropertiesEXT
*)ext
;
1667 /* From the SKL PRM Vol. 2d, docs for RENDER_SURFACE_STATE::Surface
1670 * "For SURFTYPE_BUFFER non-rendertarget surfaces, this field
1671 * specifies the base address of the first element of the surface,
1672 * computed in software by adding the surface base address to the
1673 * byte offset of the element in the buffer. The base address must
1674 * be aligned to element size."
1676 * The typed dataport messages require that things be texel aligned.
1677 * Otherwise, we may just load/store the wrong data or, in the worst
1678 * case, there may be hangs.
1680 props
->storageTexelBufferOffsetAlignmentBytes
= 16;
1681 props
->storageTexelBufferOffsetSingleTexelAlignment
= true;
1683 /* The sampler, however, is much more forgiving and it can handle
1684 * arbitrary byte alignment for linear and buffer surfaces. It's
1685 * hard to find a good PRM citation for this but years of empirical
1686 * experience demonstrate that this is true.
1688 props
->uniformTexelBufferOffsetAlignmentBytes
= 1;
1689 props
->uniformTexelBufferOffsetSingleTexelAlignment
= false;
1693 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TRANSFORM_FEEDBACK_PROPERTIES_EXT
: {
1694 VkPhysicalDeviceTransformFeedbackPropertiesEXT
*props
=
1695 (VkPhysicalDeviceTransformFeedbackPropertiesEXT
*)ext
;
1697 props
->maxTransformFeedbackStreams
= MAX_XFB_STREAMS
;
1698 props
->maxTransformFeedbackBuffers
= MAX_XFB_BUFFERS
;
1699 props
->maxTransformFeedbackBufferSize
= (1ull << 32);
1700 props
->maxTransformFeedbackStreamDataSize
= 128 * 4;
1701 props
->maxTransformFeedbackBufferDataSize
= 128 * 4;
1702 props
->maxTransformFeedbackBufferDataStride
= 2048;
1703 props
->transformFeedbackQueries
= true;
1704 props
->transformFeedbackStreamsLinesTriangles
= false;
1705 props
->transformFeedbackRasterizationStreamSelect
= false;
1706 props
->transformFeedbackDraw
= true;
1710 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VERTEX_ATTRIBUTE_DIVISOR_PROPERTIES_EXT
: {
1711 VkPhysicalDeviceVertexAttributeDivisorPropertiesEXT
*props
=
1712 (VkPhysicalDeviceVertexAttributeDivisorPropertiesEXT
*)ext
;
1713 /* We have to restrict this a bit for multiview */
1714 props
->maxVertexAttribDivisor
= UINT32_MAX
/ 16;
1719 anv_debug_ignored_stype(ext
->sType
);
1725 /* We support exactly one queue family. */
1726 static const VkQueueFamilyProperties
1727 anv_queue_family_properties
= {
1728 .queueFlags
= VK_QUEUE_GRAPHICS_BIT
|
1729 VK_QUEUE_COMPUTE_BIT
|
1730 VK_QUEUE_TRANSFER_BIT
,
1732 .timestampValidBits
= 36, /* XXX: Real value here */
1733 .minImageTransferGranularity
= { 1, 1, 1 },
1736 void anv_GetPhysicalDeviceQueueFamilyProperties(
1737 VkPhysicalDevice physicalDevice
,
1739 VkQueueFamilyProperties
* pQueueFamilyProperties
)
1741 VK_OUTARRAY_MAKE(out
, pQueueFamilyProperties
, pCount
);
1743 vk_outarray_append(&out
, p
) {
1744 *p
= anv_queue_family_properties
;
1748 void anv_GetPhysicalDeviceQueueFamilyProperties2(
1749 VkPhysicalDevice physicalDevice
,
1750 uint32_t* pQueueFamilyPropertyCount
,
1751 VkQueueFamilyProperties2
* pQueueFamilyProperties
)
1754 VK_OUTARRAY_MAKE(out
, pQueueFamilyProperties
, pQueueFamilyPropertyCount
);
1756 vk_outarray_append(&out
, p
) {
1757 p
->queueFamilyProperties
= anv_queue_family_properties
;
1759 vk_foreach_struct(s
, p
->pNext
) {
1760 anv_debug_ignored_stype(s
->sType
);
1765 void anv_GetPhysicalDeviceMemoryProperties(
1766 VkPhysicalDevice physicalDevice
,
1767 VkPhysicalDeviceMemoryProperties
* pMemoryProperties
)
1769 ANV_FROM_HANDLE(anv_physical_device
, physical_device
, physicalDevice
);
1771 pMemoryProperties
->memoryTypeCount
= physical_device
->memory
.type_count
;
1772 for (uint32_t i
= 0; i
< physical_device
->memory
.type_count
; i
++) {
1773 pMemoryProperties
->memoryTypes
[i
] = (VkMemoryType
) {
1774 .propertyFlags
= physical_device
->memory
.types
[i
].propertyFlags
,
1775 .heapIndex
= physical_device
->memory
.types
[i
].heapIndex
,
1779 pMemoryProperties
->memoryHeapCount
= physical_device
->memory
.heap_count
;
1780 for (uint32_t i
= 0; i
< physical_device
->memory
.heap_count
; i
++) {
1781 pMemoryProperties
->memoryHeaps
[i
] = (VkMemoryHeap
) {
1782 .size
= physical_device
->memory
.heaps
[i
].size
,
1783 .flags
= physical_device
->memory
.heaps
[i
].flags
,
1789 anv_get_memory_budget(VkPhysicalDevice physicalDevice
,
1790 VkPhysicalDeviceMemoryBudgetPropertiesEXT
*memoryBudget
)
1792 ANV_FROM_HANDLE(anv_physical_device
, device
, physicalDevice
);
1793 uint64_t sys_available
= get_available_system_memory();
1794 assert(sys_available
> 0);
1796 VkDeviceSize total_heaps_size
= 0;
1797 for (size_t i
= 0; i
< device
->memory
.heap_count
; i
++)
1798 total_heaps_size
+= device
->memory
.heaps
[i
].size
;
1800 for (size_t i
= 0; i
< device
->memory
.heap_count
; i
++) {
1801 VkDeviceSize heap_size
= device
->memory
.heaps
[i
].size
;
1802 VkDeviceSize heap_used
= device
->memory
.heaps
[i
].used
;
1803 VkDeviceSize heap_budget
;
1805 double heap_proportion
= (double) heap_size
/ total_heaps_size
;
1806 VkDeviceSize sys_available_prop
= sys_available
* heap_proportion
;
1809 * Let's not incite the app to starve the system: report at most 90% of
1810 * available system memory.
1812 uint64_t heap_available
= sys_available_prop
* 9 / 10;
1813 heap_budget
= MIN2(heap_size
, heap_used
+ heap_available
);
1816 * Round down to the nearest MB
1818 heap_budget
&= ~((1ull << 20) - 1);
1821 * The heapBudget value must be non-zero for array elements less than
1822 * VkPhysicalDeviceMemoryProperties::memoryHeapCount. The heapBudget
1823 * value must be less than or equal to VkMemoryHeap::size for each heap.
1825 assert(0 < heap_budget
&& heap_budget
<= heap_size
);
1827 memoryBudget
->heapUsage
[i
] = heap_used
;
1828 memoryBudget
->heapBudget
[i
] = heap_budget
;
1831 /* The heapBudget and heapUsage values must be zero for array elements
1832 * greater than or equal to VkPhysicalDeviceMemoryProperties::memoryHeapCount
1834 for (uint32_t i
= device
->memory
.heap_count
; i
< VK_MAX_MEMORY_HEAPS
; i
++) {
1835 memoryBudget
->heapBudget
[i
] = 0;
1836 memoryBudget
->heapUsage
[i
] = 0;
1840 void anv_GetPhysicalDeviceMemoryProperties2(
1841 VkPhysicalDevice physicalDevice
,
1842 VkPhysicalDeviceMemoryProperties2
* pMemoryProperties
)
1844 anv_GetPhysicalDeviceMemoryProperties(physicalDevice
,
1845 &pMemoryProperties
->memoryProperties
);
1847 vk_foreach_struct(ext
, pMemoryProperties
->pNext
) {
1848 switch (ext
->sType
) {
1849 case VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_MEMORY_BUDGET_PROPERTIES_EXT
:
1850 anv_get_memory_budget(physicalDevice
, (void*)ext
);
1853 anv_debug_ignored_stype(ext
->sType
);
1860 anv_GetDeviceGroupPeerMemoryFeatures(
1863 uint32_t localDeviceIndex
,
1864 uint32_t remoteDeviceIndex
,
1865 VkPeerMemoryFeatureFlags
* pPeerMemoryFeatures
)
1867 assert(localDeviceIndex
== 0 && remoteDeviceIndex
== 0);
1868 *pPeerMemoryFeatures
= VK_PEER_MEMORY_FEATURE_COPY_SRC_BIT
|
1869 VK_PEER_MEMORY_FEATURE_COPY_DST_BIT
|
1870 VK_PEER_MEMORY_FEATURE_GENERIC_SRC_BIT
|
1871 VK_PEER_MEMORY_FEATURE_GENERIC_DST_BIT
;
1874 PFN_vkVoidFunction
anv_GetInstanceProcAddr(
1875 VkInstance _instance
,
1878 ANV_FROM_HANDLE(anv_instance
, instance
, _instance
);
1880 /* The Vulkan 1.0 spec for vkGetInstanceProcAddr has a table of exactly
1881 * when we have to return valid function pointers, NULL, or it's left
1882 * undefined. See the table for exact details.
1887 #define LOOKUP_ANV_ENTRYPOINT(entrypoint) \
1888 if (strcmp(pName, "vk" #entrypoint) == 0) \
1889 return (PFN_vkVoidFunction)anv_##entrypoint
1891 LOOKUP_ANV_ENTRYPOINT(EnumerateInstanceExtensionProperties
);
1892 LOOKUP_ANV_ENTRYPOINT(EnumerateInstanceLayerProperties
);
1893 LOOKUP_ANV_ENTRYPOINT(EnumerateInstanceVersion
);
1894 LOOKUP_ANV_ENTRYPOINT(CreateInstance
);
1896 #undef LOOKUP_ANV_ENTRYPOINT
1898 if (instance
== NULL
)
1901 int idx
= anv_get_instance_entrypoint_index(pName
);
1903 return instance
->dispatch
.entrypoints
[idx
];
1905 idx
= anv_get_device_entrypoint_index(pName
);
1907 return instance
->device_dispatch
.entrypoints
[idx
];
1912 /* With version 1+ of the loader interface the ICD should expose
1913 * vk_icdGetInstanceProcAddr to work around certain LD_PRELOAD issues seen in apps.
1916 VKAPI_ATTR PFN_vkVoidFunction VKAPI_CALL
vk_icdGetInstanceProcAddr(
1917 VkInstance instance
,
1921 VKAPI_ATTR PFN_vkVoidFunction VKAPI_CALL
vk_icdGetInstanceProcAddr(
1922 VkInstance instance
,
1925 return anv_GetInstanceProcAddr(instance
, pName
);
1928 PFN_vkVoidFunction
anv_GetDeviceProcAddr(
1932 ANV_FROM_HANDLE(anv_device
, device
, _device
);
1934 if (!device
|| !pName
)
1937 int idx
= anv_get_device_entrypoint_index(pName
);
1941 return device
->dispatch
.entrypoints
[idx
];
1945 anv_CreateDebugReportCallbackEXT(VkInstance _instance
,
1946 const VkDebugReportCallbackCreateInfoEXT
* pCreateInfo
,
1947 const VkAllocationCallbacks
* pAllocator
,
1948 VkDebugReportCallbackEXT
* pCallback
)
1950 ANV_FROM_HANDLE(anv_instance
, instance
, _instance
);
1951 return vk_create_debug_report_callback(&instance
->debug_report_callbacks
,
1952 pCreateInfo
, pAllocator
, &instance
->alloc
,
1957 anv_DestroyDebugReportCallbackEXT(VkInstance _instance
,
1958 VkDebugReportCallbackEXT _callback
,
1959 const VkAllocationCallbacks
* pAllocator
)
1961 ANV_FROM_HANDLE(anv_instance
, instance
, _instance
);
1962 vk_destroy_debug_report_callback(&instance
->debug_report_callbacks
,
1963 _callback
, pAllocator
, &instance
->alloc
);
1967 anv_DebugReportMessageEXT(VkInstance _instance
,
1968 VkDebugReportFlagsEXT flags
,
1969 VkDebugReportObjectTypeEXT objectType
,
1972 int32_t messageCode
,
1973 const char* pLayerPrefix
,
1974 const char* pMessage
)
1976 ANV_FROM_HANDLE(anv_instance
, instance
, _instance
);
1977 vk_debug_report(&instance
->debug_report_callbacks
, flags
, objectType
,
1978 object
, location
, messageCode
, pLayerPrefix
, pMessage
);
1982 anv_queue_init(struct anv_device
*device
, struct anv_queue
*queue
)
1984 queue
->_loader_data
.loaderMagic
= ICD_LOADER_MAGIC
;
1985 queue
->device
= device
;
1990 anv_queue_finish(struct anv_queue
*queue
)
1994 static struct anv_state
1995 anv_state_pool_emit_data(struct anv_state_pool
*pool
, size_t size
, size_t align
, const void *p
)
1997 struct anv_state state
;
1999 state
= anv_state_pool_alloc(pool
, size
, align
);
2000 memcpy(state
.map
, p
, size
);
2005 /* Haswell border color is a bit of a disaster. Float and unorm formats use a
2006 * straightforward 32-bit float color in the first 64 bytes. Instead of using
2007 * a nice float/integer union like Gen8+, Haswell specifies the integer border
2008 * color as a separate entry /after/ the float color. The layout of this entry
2009 * also depends on the format's bpp (with extra hacks for RG32), and overlaps.
2011 * Since we don't know the format/bpp, we can't make any of the border colors
2012 * containing '1' work for all formats, as it would be in the wrong place for
2013 * some of them. We opt to make 32-bit integers work as this seems like the
2014 * most common option. Fortunately, transparent black works regardless, as
2015 * all zeroes is the same in every bit-size.
2017 struct hsw_border_color
{
2021 uint32_t _pad1
[108];
2024 struct gen8_border_color
{
2029 /* Pad out to 64 bytes */
2034 anv_device_init_border_colors(struct anv_device
*device
)
2036 if (device
->info
.is_haswell
) {
2037 static const struct hsw_border_color border_colors
[] = {
2038 [VK_BORDER_COLOR_FLOAT_TRANSPARENT_BLACK
] = { .float32
= { 0.0, 0.0, 0.0, 0.0 } },
2039 [VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK
] = { .float32
= { 0.0, 0.0, 0.0, 1.0 } },
2040 [VK_BORDER_COLOR_FLOAT_OPAQUE_WHITE
] = { .float32
= { 1.0, 1.0, 1.0, 1.0 } },
2041 [VK_BORDER_COLOR_INT_TRANSPARENT_BLACK
] = { .uint32
= { 0, 0, 0, 0 } },
2042 [VK_BORDER_COLOR_INT_OPAQUE_BLACK
] = { .uint32
= { 0, 0, 0, 1 } },
2043 [VK_BORDER_COLOR_INT_OPAQUE_WHITE
] = { .uint32
= { 1, 1, 1, 1 } },
2046 device
->border_colors
=
2047 anv_state_pool_emit_data(&device
->dynamic_state_pool
,
2048 sizeof(border_colors
), 512, border_colors
);
2050 static const struct gen8_border_color border_colors
[] = {
2051 [VK_BORDER_COLOR_FLOAT_TRANSPARENT_BLACK
] = { .float32
= { 0.0, 0.0, 0.0, 0.0 } },
2052 [VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK
] = { .float32
= { 0.0, 0.0, 0.0, 1.0 } },
2053 [VK_BORDER_COLOR_FLOAT_OPAQUE_WHITE
] = { .float32
= { 1.0, 1.0, 1.0, 1.0 } },
2054 [VK_BORDER_COLOR_INT_TRANSPARENT_BLACK
] = { .uint32
= { 0, 0, 0, 0 } },
2055 [VK_BORDER_COLOR_INT_OPAQUE_BLACK
] = { .uint32
= { 0, 0, 0, 1 } },
2056 [VK_BORDER_COLOR_INT_OPAQUE_WHITE
] = { .uint32
= { 1, 1, 1, 1 } },
2059 device
->border_colors
=
2060 anv_state_pool_emit_data(&device
->dynamic_state_pool
,
2061 sizeof(border_colors
), 64, border_colors
);
2066 anv_device_init_trivial_batch(struct anv_device
*device
)
2068 anv_bo_init_new(&device
->trivial_batch_bo
, device
, 4096);
2070 if (device
->instance
->physicalDevice
.has_exec_async
)
2071 device
->trivial_batch_bo
.flags
|= EXEC_OBJECT_ASYNC
;
2073 if (device
->instance
->physicalDevice
.use_softpin
)
2074 device
->trivial_batch_bo
.flags
|= EXEC_OBJECT_PINNED
;
2076 anv_vma_alloc(device
, &device
->trivial_batch_bo
);
2078 void *map
= anv_gem_mmap(device
, device
->trivial_batch_bo
.gem_handle
,
2081 struct anv_batch batch
= {
2087 anv_batch_emit(&batch
, GEN7_MI_BATCH_BUFFER_END
, bbe
);
2088 anv_batch_emit(&batch
, GEN7_MI_NOOP
, noop
);
2090 if (!device
->info
.has_llc
)
2091 gen_clflush_range(map
, batch
.next
- map
);
2093 anv_gem_munmap(map
, device
->trivial_batch_bo
.size
);
2096 VkResult
anv_EnumerateDeviceExtensionProperties(
2097 VkPhysicalDevice physicalDevice
,
2098 const char* pLayerName
,
2099 uint32_t* pPropertyCount
,
2100 VkExtensionProperties
* pProperties
)
2102 ANV_FROM_HANDLE(anv_physical_device
, device
, physicalDevice
);
2103 VK_OUTARRAY_MAKE(out
, pProperties
, pPropertyCount
);
2105 for (int i
= 0; i
< ANV_DEVICE_EXTENSION_COUNT
; i
++) {
2106 if (device
->supported_extensions
.extensions
[i
]) {
2107 vk_outarray_append(&out
, prop
) {
2108 *prop
= anv_device_extensions
[i
];
2113 return vk_outarray_status(&out
);
2117 anv_device_init_dispatch(struct anv_device
*device
)
2119 const struct anv_device_dispatch_table
*genX_table
;
2120 switch (device
->info
.gen
) {
2122 genX_table
= &gen11_device_dispatch_table
;
2125 genX_table
= &gen10_device_dispatch_table
;
2128 genX_table
= &gen9_device_dispatch_table
;
2131 genX_table
= &gen8_device_dispatch_table
;
2134 if (device
->info
.is_haswell
)
2135 genX_table
= &gen75_device_dispatch_table
;
2137 genX_table
= &gen7_device_dispatch_table
;
2140 unreachable("unsupported gen\n");
2143 for (unsigned i
= 0; i
< ARRAY_SIZE(device
->dispatch
.entrypoints
); i
++) {
2144 /* Vulkan requires that entrypoints for extensions which have not been
2145 * enabled must not be advertised.
2147 if (!anv_device_entrypoint_is_enabled(i
, device
->instance
->app_info
.api_version
,
2148 &device
->instance
->enabled_extensions
,
2149 &device
->enabled_extensions
)) {
2150 device
->dispatch
.entrypoints
[i
] = NULL
;
2151 } else if (genX_table
->entrypoints
[i
]) {
2152 device
->dispatch
.entrypoints
[i
] = genX_table
->entrypoints
[i
];
2154 device
->dispatch
.entrypoints
[i
] =
2155 anv_device_dispatch_table
.entrypoints
[i
];
2161 vk_priority_to_gen(int priority
)
2164 case VK_QUEUE_GLOBAL_PRIORITY_LOW_EXT
:
2165 return GEN_CONTEXT_LOW_PRIORITY
;
2166 case VK_QUEUE_GLOBAL_PRIORITY_MEDIUM_EXT
:
2167 return GEN_CONTEXT_MEDIUM_PRIORITY
;
2168 case VK_QUEUE_GLOBAL_PRIORITY_HIGH_EXT
:
2169 return GEN_CONTEXT_HIGH_PRIORITY
;
2170 case VK_QUEUE_GLOBAL_PRIORITY_REALTIME_EXT
:
2171 return GEN_CONTEXT_REALTIME_PRIORITY
;
2173 unreachable("Invalid priority");
2178 anv_device_init_hiz_clear_value_bo(struct anv_device
*device
)
2180 anv_bo_init_new(&device
->hiz_clear_bo
, device
, 4096);
2182 if (device
->instance
->physicalDevice
.has_exec_async
)
2183 device
->hiz_clear_bo
.flags
|= EXEC_OBJECT_ASYNC
;
2185 if (device
->instance
->physicalDevice
.use_softpin
)
2186 device
->hiz_clear_bo
.flags
|= EXEC_OBJECT_PINNED
;
2188 anv_vma_alloc(device
, &device
->hiz_clear_bo
);
2190 uint32_t *map
= anv_gem_mmap(device
, device
->hiz_clear_bo
.gem_handle
,
2193 union isl_color_value hiz_clear
= { .u32
= { 0, } };
2194 hiz_clear
.f32
[0] = ANV_HZ_FC_VAL
;
2196 memcpy(map
, hiz_clear
.u32
, sizeof(hiz_clear
.u32
));
2197 anv_gem_munmap(map
, device
->hiz_clear_bo
.size
);
2201 get_bo_from_pool(struct gen_batch_decode_bo
*ret
,
2202 struct anv_block_pool
*pool
,
2205 for (uint32_t i
= 0; i
< pool
->nbos
; i
++) {
2206 uint64_t bo_address
= pool
->bos
[i
].offset
& (~0ull >> 16);
2207 uint32_t bo_size
= pool
->bos
[i
].size
;
2208 if (address
>= bo_address
&& address
< (bo_address
+ bo_size
)) {
2209 *ret
= (struct gen_batch_decode_bo
) {
2212 .map
= pool
->bos
[i
].map
,
2220 /* Finding a buffer for batch decoding */
2221 static struct gen_batch_decode_bo
2222 decode_get_bo(void *v_batch
, bool ppgtt
, uint64_t address
)
2224 struct anv_device
*device
= v_batch
;
2225 struct gen_batch_decode_bo ret_bo
= {};
2229 if (get_bo_from_pool(&ret_bo
, &device
->dynamic_state_pool
.block_pool
, address
))
2231 if (get_bo_from_pool(&ret_bo
, &device
->instruction_state_pool
.block_pool
, address
))
2233 if (get_bo_from_pool(&ret_bo
, &device
->binding_table_pool
.block_pool
, address
))
2235 if (get_bo_from_pool(&ret_bo
, &device
->surface_state_pool
.block_pool
, address
))
2238 if (!device
->cmd_buffer_being_decoded
)
2239 return (struct gen_batch_decode_bo
) { };
2241 struct anv_batch_bo
**bo
;
2243 u_vector_foreach(bo
, &device
->cmd_buffer_being_decoded
->seen_bbos
) {
2244 /* The decoder zeroes out the top 16 bits, so we need to as well */
2245 uint64_t bo_address
= (*bo
)->bo
.offset
& (~0ull >> 16);
2247 if (address
>= bo_address
&& address
< bo_address
+ (*bo
)->bo
.size
) {
2248 return (struct gen_batch_decode_bo
) {
2250 .size
= (*bo
)->bo
.size
,
2251 .map
= (*bo
)->bo
.map
,
2256 return (struct gen_batch_decode_bo
) { };
2259 VkResult
anv_CreateDevice(
2260 VkPhysicalDevice physicalDevice
,
2261 const VkDeviceCreateInfo
* pCreateInfo
,
2262 const VkAllocationCallbacks
* pAllocator
,
2265 ANV_FROM_HANDLE(anv_physical_device
, physical_device
, physicalDevice
);
2267 struct anv_device
*device
;
2269 assert(pCreateInfo
->sType
== VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO
);
2271 struct anv_device_extension_table enabled_extensions
= { };
2272 for (uint32_t i
= 0; i
< pCreateInfo
->enabledExtensionCount
; i
++) {
2274 for (idx
= 0; idx
< ANV_DEVICE_EXTENSION_COUNT
; idx
++) {
2275 if (strcmp(pCreateInfo
->ppEnabledExtensionNames
[i
],
2276 anv_device_extensions
[idx
].extensionName
) == 0)
2280 if (idx
>= ANV_DEVICE_EXTENSION_COUNT
)
2281 return vk_error(VK_ERROR_EXTENSION_NOT_PRESENT
);
2283 if (!physical_device
->supported_extensions
.extensions
[idx
])
2284 return vk_error(VK_ERROR_EXTENSION_NOT_PRESENT
);
2286 enabled_extensions
.extensions
[idx
] = true;
2289 /* Check enabled features */
2290 if (pCreateInfo
->pEnabledFeatures
) {
2291 VkPhysicalDeviceFeatures supported_features
;
2292 anv_GetPhysicalDeviceFeatures(physicalDevice
, &supported_features
);
2293 VkBool32
*supported_feature
= (VkBool32
*)&supported_features
;
2294 VkBool32
*enabled_feature
= (VkBool32
*)pCreateInfo
->pEnabledFeatures
;
2295 unsigned num_features
= sizeof(VkPhysicalDeviceFeatures
) / sizeof(VkBool32
);
2296 for (uint32_t i
= 0; i
< num_features
; i
++) {
2297 if (enabled_feature
[i
] && !supported_feature
[i
])
2298 return vk_error(VK_ERROR_FEATURE_NOT_PRESENT
);
2302 /* Check requested queues and fail if we are requested to create any
2303 * queues with flags we don't support.
2305 assert(pCreateInfo
->queueCreateInfoCount
> 0);
2306 for (uint32_t i
= 0; i
< pCreateInfo
->queueCreateInfoCount
; i
++) {
2307 if (pCreateInfo
->pQueueCreateInfos
[i
].flags
!= 0)
2308 return vk_error(VK_ERROR_INITIALIZATION_FAILED
);
2311 /* Check if client specified queue priority. */
2312 const VkDeviceQueueGlobalPriorityCreateInfoEXT
*queue_priority
=
2313 vk_find_struct_const(pCreateInfo
->pQueueCreateInfos
[0].pNext
,
2314 DEVICE_QUEUE_GLOBAL_PRIORITY_CREATE_INFO_EXT
);
2316 VkQueueGlobalPriorityEXT priority
=
2317 queue_priority
? queue_priority
->globalPriority
:
2318 VK_QUEUE_GLOBAL_PRIORITY_MEDIUM_EXT
;
2320 device
= vk_alloc2(&physical_device
->instance
->alloc
, pAllocator
,
2322 VK_SYSTEM_ALLOCATION_SCOPE_DEVICE
);
2324 return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY
);
2326 if (INTEL_DEBUG
& DEBUG_BATCH
) {
2327 const unsigned decode_flags
=
2328 GEN_BATCH_DECODE_FULL
|
2329 ((INTEL_DEBUG
& DEBUG_COLOR
) ? GEN_BATCH_DECODE_IN_COLOR
: 0) |
2330 GEN_BATCH_DECODE_OFFSETS
|
2331 GEN_BATCH_DECODE_FLOATS
;
2333 gen_batch_decode_ctx_init(&device
->decoder_ctx
,
2334 &physical_device
->info
,
2335 stderr
, decode_flags
, NULL
,
2336 decode_get_bo
, NULL
, device
);
2339 device
->_loader_data
.loaderMagic
= ICD_LOADER_MAGIC
;
2340 device
->instance
= physical_device
->instance
;
2341 device
->chipset_id
= physical_device
->chipset_id
;
2342 device
->no_hw
= physical_device
->no_hw
;
2343 device
->_lost
= false;
2346 device
->alloc
= *pAllocator
;
2348 device
->alloc
= physical_device
->instance
->alloc
;
2350 /* XXX(chadv): Can we dup() physicalDevice->fd here? */
2351 device
->fd
= open(physical_device
->path
, O_RDWR
| O_CLOEXEC
);
2352 if (device
->fd
== -1) {
2353 result
= vk_error(VK_ERROR_INITIALIZATION_FAILED
);
2357 device
->context_id
= anv_gem_create_context(device
);
2358 if (device
->context_id
== -1) {
2359 result
= vk_error(VK_ERROR_INITIALIZATION_FAILED
);
2363 if (physical_device
->use_softpin
) {
2364 if (pthread_mutex_init(&device
->vma_mutex
, NULL
) != 0) {
2365 result
= vk_error(VK_ERROR_INITIALIZATION_FAILED
);
2369 /* keep the page with address zero out of the allocator */
2370 struct anv_memory_heap
*low_heap
=
2371 &physical_device
->memory
.heaps
[physical_device
->memory
.heap_count
- 1];
2372 util_vma_heap_init(&device
->vma_lo
, low_heap
->vma_start
, low_heap
->vma_size
);
2373 device
->vma_lo_available
= low_heap
->size
;
2375 struct anv_memory_heap
*high_heap
=
2376 &physical_device
->memory
.heaps
[0];
2377 util_vma_heap_init(&device
->vma_hi
, high_heap
->vma_start
, high_heap
->vma_size
);
2378 device
->vma_hi_available
= physical_device
->memory
.heap_count
== 1 ? 0 :
2382 list_inithead(&device
->memory_objects
);
2384 /* As per spec, the driver implementation may deny requests to acquire
2385 * a priority above the default priority (MEDIUM) if the caller does not
2386 * have sufficient privileges. In this scenario VK_ERROR_NOT_PERMITTED_EXT
2389 if (physical_device
->has_context_priority
) {
2390 int err
= anv_gem_set_context_param(device
->fd
, device
->context_id
,
2391 I915_CONTEXT_PARAM_PRIORITY
,
2392 vk_priority_to_gen(priority
));
2393 if (err
!= 0 && priority
> VK_QUEUE_GLOBAL_PRIORITY_MEDIUM_EXT
) {
2394 result
= vk_error(VK_ERROR_NOT_PERMITTED_EXT
);
2399 device
->info
= physical_device
->info
;
2400 device
->isl_dev
= physical_device
->isl_dev
;
2402 /* On Broadwell and later, we can use batch chaining to more efficiently
2403 * implement growing command buffers. Prior to Haswell, the kernel
2404 * command parser gets in the way and we have to fall back to growing
2407 device
->can_chain_batches
= device
->info
.gen
>= 8;
2409 device
->robust_buffer_access
= pCreateInfo
->pEnabledFeatures
&&
2410 pCreateInfo
->pEnabledFeatures
->robustBufferAccess
;
2411 device
->enabled_extensions
= enabled_extensions
;
2413 anv_device_init_dispatch(device
);
2415 if (pthread_mutex_init(&device
->mutex
, NULL
) != 0) {
2416 result
= vk_error(VK_ERROR_INITIALIZATION_FAILED
);
2417 goto fail_context_id
;
2420 pthread_condattr_t condattr
;
2421 if (pthread_condattr_init(&condattr
) != 0) {
2422 result
= vk_error(VK_ERROR_INITIALIZATION_FAILED
);
2425 if (pthread_condattr_setclock(&condattr
, CLOCK_MONOTONIC
) != 0) {
2426 pthread_condattr_destroy(&condattr
);
2427 result
= vk_error(VK_ERROR_INITIALIZATION_FAILED
);
2430 if (pthread_cond_init(&device
->queue_submit
, &condattr
) != 0) {
2431 pthread_condattr_destroy(&condattr
);
2432 result
= vk_error(VK_ERROR_INITIALIZATION_FAILED
);
2435 pthread_condattr_destroy(&condattr
);
2438 (physical_device
->supports_48bit_addresses
? EXEC_OBJECT_SUPPORTS_48B_ADDRESS
: 0) |
2439 (physical_device
->has_exec_async
? EXEC_OBJECT_ASYNC
: 0) |
2440 (physical_device
->has_exec_capture
? EXEC_OBJECT_CAPTURE
: 0) |
2441 (physical_device
->use_softpin
? EXEC_OBJECT_PINNED
: 0);
2443 anv_bo_pool_init(&device
->batch_bo_pool
, device
, bo_flags
);
2445 result
= anv_bo_cache_init(&device
->bo_cache
);
2446 if (result
!= VK_SUCCESS
)
2447 goto fail_batch_bo_pool
;
2449 if (!physical_device
->use_softpin
)
2450 bo_flags
&= ~EXEC_OBJECT_SUPPORTS_48B_ADDRESS
;
2452 result
= anv_state_pool_init(&device
->dynamic_state_pool
, device
,
2453 DYNAMIC_STATE_POOL_MIN_ADDRESS
,
2456 if (result
!= VK_SUCCESS
)
2459 result
= anv_state_pool_init(&device
->instruction_state_pool
, device
,
2460 INSTRUCTION_STATE_POOL_MIN_ADDRESS
,
2463 if (result
!= VK_SUCCESS
)
2464 goto fail_dynamic_state_pool
;
2466 result
= anv_state_pool_init(&device
->surface_state_pool
, device
,
2467 SURFACE_STATE_POOL_MIN_ADDRESS
,
2470 if (result
!= VK_SUCCESS
)
2471 goto fail_instruction_state_pool
;
2473 if (physical_device
->use_softpin
) {
2474 result
= anv_state_pool_init(&device
->binding_table_pool
, device
,
2475 BINDING_TABLE_POOL_MIN_ADDRESS
,
2478 if (result
!= VK_SUCCESS
)
2479 goto fail_surface_state_pool
;
2482 result
= anv_bo_init_new(&device
->workaround_bo
, device
, 4096);
2483 if (result
!= VK_SUCCESS
)
2484 goto fail_binding_table_pool
;
2486 if (physical_device
->use_softpin
)
2487 device
->workaround_bo
.flags
|= EXEC_OBJECT_PINNED
;
2489 if (!anv_vma_alloc(device
, &device
->workaround_bo
))
2490 goto fail_workaround_bo
;
2492 anv_device_init_trivial_batch(device
);
2494 if (device
->info
.gen
>= 10)
2495 anv_device_init_hiz_clear_value_bo(device
);
2497 anv_scratch_pool_init(device
, &device
->scratch_pool
);
2499 anv_queue_init(device
, &device
->queue
);
2501 switch (device
->info
.gen
) {
2503 if (!device
->info
.is_haswell
)
2504 result
= gen7_init_device_state(device
);
2506 result
= gen75_init_device_state(device
);
2509 result
= gen8_init_device_state(device
);
2512 result
= gen9_init_device_state(device
);
2515 result
= gen10_init_device_state(device
);
2518 result
= gen11_init_device_state(device
);
2521 /* Shouldn't get here as we don't create physical devices for any other
2523 unreachable("unhandled gen");
2525 if (result
!= VK_SUCCESS
)
2526 goto fail_workaround_bo
;
2528 anv_pipeline_cache_init(&device
->default_pipeline_cache
, device
, true);
2530 anv_device_init_blorp(device
);
2532 anv_device_init_border_colors(device
);
2534 *pDevice
= anv_device_to_handle(device
);
2539 anv_queue_finish(&device
->queue
);
2540 anv_scratch_pool_finish(device
, &device
->scratch_pool
);
2541 anv_gem_munmap(device
->workaround_bo
.map
, device
->workaround_bo
.size
);
2542 anv_gem_close(device
, device
->workaround_bo
.gem_handle
);
2543 fail_binding_table_pool
:
2544 if (physical_device
->use_softpin
)
2545 anv_state_pool_finish(&device
->binding_table_pool
);
2546 fail_surface_state_pool
:
2547 anv_state_pool_finish(&device
->surface_state_pool
);
2548 fail_instruction_state_pool
:
2549 anv_state_pool_finish(&device
->instruction_state_pool
);
2550 fail_dynamic_state_pool
:
2551 anv_state_pool_finish(&device
->dynamic_state_pool
);
2553 anv_bo_cache_finish(&device
->bo_cache
);
2555 anv_bo_pool_finish(&device
->batch_bo_pool
);
2556 pthread_cond_destroy(&device
->queue_submit
);
2558 pthread_mutex_destroy(&device
->mutex
);
2560 anv_gem_destroy_context(device
, device
->context_id
);
2564 vk_free(&device
->alloc
, device
);
2569 void anv_DestroyDevice(
2571 const VkAllocationCallbacks
* pAllocator
)
2573 ANV_FROM_HANDLE(anv_device
, device
, _device
);
2574 struct anv_physical_device
*physical_device
;
2579 physical_device
= &device
->instance
->physicalDevice
;
2581 anv_device_finish_blorp(device
);
2583 anv_pipeline_cache_finish(&device
->default_pipeline_cache
);
2585 anv_queue_finish(&device
->queue
);
2587 #ifdef HAVE_VALGRIND
2588 /* We only need to free these to prevent valgrind errors. The backing
2589 * BO will go away in a couple of lines so we don't actually leak.
2591 anv_state_pool_free(&device
->dynamic_state_pool
, device
->border_colors
);
2594 anv_scratch_pool_finish(device
, &device
->scratch_pool
);
2596 anv_gem_munmap(device
->workaround_bo
.map
, device
->workaround_bo
.size
);
2597 anv_vma_free(device
, &device
->workaround_bo
);
2598 anv_gem_close(device
, device
->workaround_bo
.gem_handle
);
2600 anv_vma_free(device
, &device
->trivial_batch_bo
);
2601 anv_gem_close(device
, device
->trivial_batch_bo
.gem_handle
);
2602 if (device
->info
.gen
>= 10)
2603 anv_gem_close(device
, device
->hiz_clear_bo
.gem_handle
);
2605 if (physical_device
->use_softpin
)
2606 anv_state_pool_finish(&device
->binding_table_pool
);
2607 anv_state_pool_finish(&device
->surface_state_pool
);
2608 anv_state_pool_finish(&device
->instruction_state_pool
);
2609 anv_state_pool_finish(&device
->dynamic_state_pool
);
2611 anv_bo_cache_finish(&device
->bo_cache
);
2613 anv_bo_pool_finish(&device
->batch_bo_pool
);
2615 pthread_cond_destroy(&device
->queue_submit
);
2616 pthread_mutex_destroy(&device
->mutex
);
2618 anv_gem_destroy_context(device
, device
->context_id
);
2620 if (INTEL_DEBUG
& DEBUG_BATCH
)
2621 gen_batch_decode_ctx_finish(&device
->decoder_ctx
);
2625 vk_free(&device
->alloc
, device
);
2628 VkResult
anv_EnumerateInstanceLayerProperties(
2629 uint32_t* pPropertyCount
,
2630 VkLayerProperties
* pProperties
)
2632 if (pProperties
== NULL
) {
2633 *pPropertyCount
= 0;
2637 /* None supported at this time */
2638 return vk_error(VK_ERROR_LAYER_NOT_PRESENT
);
2641 VkResult
anv_EnumerateDeviceLayerProperties(
2642 VkPhysicalDevice physicalDevice
,
2643 uint32_t* pPropertyCount
,
2644 VkLayerProperties
* pProperties
)
2646 if (pProperties
== NULL
) {
2647 *pPropertyCount
= 0;
2651 /* None supported at this time */
2652 return vk_error(VK_ERROR_LAYER_NOT_PRESENT
);
2655 void anv_GetDeviceQueue(
2657 uint32_t queueNodeIndex
,
2658 uint32_t queueIndex
,
2661 ANV_FROM_HANDLE(anv_device
, device
, _device
);
2663 assert(queueIndex
== 0);
2665 *pQueue
= anv_queue_to_handle(&device
->queue
);
2668 void anv_GetDeviceQueue2(
2670 const VkDeviceQueueInfo2
* pQueueInfo
,
2673 ANV_FROM_HANDLE(anv_device
, device
, _device
);
2675 assert(pQueueInfo
->queueIndex
== 0);
2677 if (pQueueInfo
->flags
== device
->queue
.flags
)
2678 *pQueue
= anv_queue_to_handle(&device
->queue
);
2684 _anv_device_set_lost(struct anv_device
*device
,
2685 const char *file
, int line
,
2686 const char *msg
, ...)
2691 device
->_lost
= true;
2694 err
= __vk_errorv(device
->instance
, device
,
2695 VK_DEBUG_REPORT_OBJECT_TYPE_DEVICE_EXT
,
2696 VK_ERROR_DEVICE_LOST
, file
, line
, msg
, ap
);
2699 if (env_var_as_boolean("ANV_ABORT_ON_DEVICE_LOSS", false))
2706 anv_device_query_status(struct anv_device
*device
)
2708 /* This isn't likely as most of the callers of this function already check
2709 * for it. However, it doesn't hurt to check and it potentially lets us
2712 if (anv_device_is_lost(device
))
2713 return VK_ERROR_DEVICE_LOST
;
2715 uint32_t active
, pending
;
2716 int ret
= anv_gem_gpu_get_reset_stats(device
, &active
, &pending
);
2718 /* We don't know the real error. */
2719 return anv_device_set_lost(device
, "get_reset_stats failed: %m");
2723 return anv_device_set_lost(device
, "GPU hung on one of our command buffers");
2724 } else if (pending
) {
2725 return anv_device_set_lost(device
, "GPU hung with commands in-flight");
2732 anv_device_bo_busy(struct anv_device
*device
, struct anv_bo
*bo
)
2734 /* Note: This only returns whether or not the BO is in use by an i915 GPU.
2735 * Other usages of the BO (such as on different hardware) will not be
2736 * flagged as "busy" by this ioctl. Use with care.
2738 int ret
= anv_gem_busy(device
, bo
->gem_handle
);
2740 return VK_NOT_READY
;
2741 } else if (ret
== -1) {
2742 /* We don't know the real error. */
2743 return anv_device_set_lost(device
, "gem wait failed: %m");
2746 /* Query for device status after the busy call. If the BO we're checking
2747 * got caught in a GPU hang we don't want to return VK_SUCCESS to the
2748 * client because it clearly doesn't have valid data. Yes, this most
2749 * likely means an ioctl, but we just did an ioctl to query the busy status
2750 * so it's no great loss.
2752 return anv_device_query_status(device
);
2756 anv_device_wait(struct anv_device
*device
, struct anv_bo
*bo
,
2759 int ret
= anv_gem_wait(device
, bo
->gem_handle
, &timeout
);
2760 if (ret
== -1 && errno
== ETIME
) {
2762 } else if (ret
== -1) {
2763 /* We don't know the real error. */
2764 return anv_device_set_lost(device
, "gem wait failed: %m");
2767 /* Query for device status after the wait. If the BO we're waiting on got
2768 * caught in a GPU hang we don't want to return VK_SUCCESS to the client
2769 * because it clearly doesn't have valid data. Yes, this most likely means
2770 * an ioctl, but we just did an ioctl to wait so it's no great loss.
2772 return anv_device_query_status(device
);
2775 VkResult
anv_DeviceWaitIdle(
2778 ANV_FROM_HANDLE(anv_device
, device
, _device
);
2779 if (anv_device_is_lost(device
))
2780 return VK_ERROR_DEVICE_LOST
;
2782 struct anv_batch batch
;
2785 batch
.start
= batch
.next
= cmds
;
2786 batch
.end
= (void *) cmds
+ sizeof(cmds
);
2788 anv_batch_emit(&batch
, GEN7_MI_BATCH_BUFFER_END
, bbe
);
2789 anv_batch_emit(&batch
, GEN7_MI_NOOP
, noop
);
2791 return anv_device_submit_simple_batch(device
, &batch
);
2795 anv_vma_alloc(struct anv_device
*device
, struct anv_bo
*bo
)
2797 if (!(bo
->flags
& EXEC_OBJECT_PINNED
))
2800 pthread_mutex_lock(&device
->vma_mutex
);
2804 if (bo
->flags
& EXEC_OBJECT_SUPPORTS_48B_ADDRESS
&&
2805 device
->vma_hi_available
>= bo
->size
) {
2806 uint64_t addr
= util_vma_heap_alloc(&device
->vma_hi
, bo
->size
, 4096);
2808 bo
->offset
= gen_canonical_address(addr
);
2809 assert(addr
== gen_48b_address(bo
->offset
));
2810 device
->vma_hi_available
-= bo
->size
;
2814 if (bo
->offset
== 0 && device
->vma_lo_available
>= bo
->size
) {
2815 uint64_t addr
= util_vma_heap_alloc(&device
->vma_lo
, bo
->size
, 4096);
2817 bo
->offset
= gen_canonical_address(addr
);
2818 assert(addr
== gen_48b_address(bo
->offset
));
2819 device
->vma_lo_available
-= bo
->size
;
2823 pthread_mutex_unlock(&device
->vma_mutex
);
2825 return bo
->offset
!= 0;
2829 anv_vma_free(struct anv_device
*device
, struct anv_bo
*bo
)
2831 if (!(bo
->flags
& EXEC_OBJECT_PINNED
))
2834 const uint64_t addr_48b
= gen_48b_address(bo
->offset
);
2836 pthread_mutex_lock(&device
->vma_mutex
);
2838 if (addr_48b
>= LOW_HEAP_MIN_ADDRESS
&&
2839 addr_48b
<= LOW_HEAP_MAX_ADDRESS
) {
2840 util_vma_heap_free(&device
->vma_lo
, addr_48b
, bo
->size
);
2841 device
->vma_lo_available
+= bo
->size
;
2843 ASSERTED
const struct anv_physical_device
*physical_device
=
2844 &device
->instance
->physicalDevice
;
2845 assert(addr_48b
>= physical_device
->memory
.heaps
[0].vma_start
&&
2846 addr_48b
< (physical_device
->memory
.heaps
[0].vma_start
+
2847 physical_device
->memory
.heaps
[0].vma_size
));
2848 util_vma_heap_free(&device
->vma_hi
, addr_48b
, bo
->size
);
2849 device
->vma_hi_available
+= bo
->size
;
2852 pthread_mutex_unlock(&device
->vma_mutex
);
2858 anv_bo_init_new(struct anv_bo
*bo
, struct anv_device
*device
, uint64_t size
)
2860 uint32_t gem_handle
= anv_gem_create(device
, size
);
2862 return vk_error(VK_ERROR_OUT_OF_DEVICE_MEMORY
);
2864 anv_bo_init(bo
, gem_handle
, size
);
2869 VkResult
anv_AllocateMemory(
2871 const VkMemoryAllocateInfo
* pAllocateInfo
,
2872 const VkAllocationCallbacks
* pAllocator
,
2873 VkDeviceMemory
* pMem
)
2875 ANV_FROM_HANDLE(anv_device
, device
, _device
);
2876 struct anv_physical_device
*pdevice
= &device
->instance
->physicalDevice
;
2877 struct anv_device_memory
*mem
;
2878 VkResult result
= VK_SUCCESS
;
2880 assert(pAllocateInfo
->sType
== VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO
);
2882 /* The Vulkan 1.0.33 spec says "allocationSize must be greater than 0". */
2883 assert(pAllocateInfo
->allocationSize
> 0);
2885 if (pAllocateInfo
->allocationSize
> MAX_MEMORY_ALLOCATION_SIZE
)
2886 return VK_ERROR_OUT_OF_DEVICE_MEMORY
;
2888 /* FINISHME: Fail if allocation request exceeds heap size. */
2890 mem
= vk_alloc2(&device
->alloc
, pAllocator
, sizeof(*mem
), 8,
2891 VK_SYSTEM_ALLOCATION_SCOPE_OBJECT
);
2893 return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY
);
2895 assert(pAllocateInfo
->memoryTypeIndex
< pdevice
->memory
.type_count
);
2896 mem
->type
= &pdevice
->memory
.types
[pAllocateInfo
->memoryTypeIndex
];
2900 mem
->host_ptr
= NULL
;
2902 uint64_t bo_flags
= 0;
2904 assert(mem
->type
->heapIndex
< pdevice
->memory
.heap_count
);
2905 if (pdevice
->memory
.heaps
[mem
->type
->heapIndex
].supports_48bit_addresses
)
2906 bo_flags
|= EXEC_OBJECT_SUPPORTS_48B_ADDRESS
;
2908 const struct wsi_memory_allocate_info
*wsi_info
=
2909 vk_find_struct_const(pAllocateInfo
->pNext
, WSI_MEMORY_ALLOCATE_INFO_MESA
);
2910 if (wsi_info
&& wsi_info
->implicit_sync
) {
2911 /* We need to set the WRITE flag on window system buffers so that GEM
2912 * will know we're writing to them and synchronize uses on other rings
2913 * (eg if the display server uses the blitter ring).
2915 bo_flags
|= EXEC_OBJECT_WRITE
;
2916 } else if (pdevice
->has_exec_async
) {
2917 bo_flags
|= EXEC_OBJECT_ASYNC
;
2920 if (pdevice
->use_softpin
)
2921 bo_flags
|= EXEC_OBJECT_PINNED
;
2923 const VkExportMemoryAllocateInfo
*export_info
=
2924 vk_find_struct_const(pAllocateInfo
->pNext
, EXPORT_MEMORY_ALLOCATE_INFO
);
2926 /* Check if we need to support Android HW buffer export. If so,
2927 * create AHardwareBuffer and import memory from it.
2929 bool android_export
= false;
2930 if (export_info
&& export_info
->handleTypes
&
2931 VK_EXTERNAL_MEMORY_HANDLE_TYPE_ANDROID_HARDWARE_BUFFER_BIT_ANDROID
)
2932 android_export
= true;
2934 /* Android memory import. */
2935 const struct VkImportAndroidHardwareBufferInfoANDROID
*ahw_import_info
=
2936 vk_find_struct_const(pAllocateInfo
->pNext
,
2937 IMPORT_ANDROID_HARDWARE_BUFFER_INFO_ANDROID
);
2939 if (ahw_import_info
) {
2940 result
= anv_import_ahw_memory(_device
, mem
, ahw_import_info
);
2941 if (result
!= VK_SUCCESS
)
2945 } else if (android_export
) {
2946 result
= anv_create_ahw_memory(_device
, mem
, pAllocateInfo
);
2947 if (result
!= VK_SUCCESS
)
2950 const struct VkImportAndroidHardwareBufferInfoANDROID import_info
= {
2953 result
= anv_import_ahw_memory(_device
, mem
, &import_info
);
2954 if (result
!= VK_SUCCESS
)
2960 const VkImportMemoryFdInfoKHR
*fd_info
=
2961 vk_find_struct_const(pAllocateInfo
->pNext
, IMPORT_MEMORY_FD_INFO_KHR
);
2963 /* The Vulkan spec permits handleType to be 0, in which case the struct is
2966 if (fd_info
&& fd_info
->handleType
) {
2967 /* At the moment, we support only the below handle types. */
2968 assert(fd_info
->handleType
==
2969 VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD_BIT
||
2970 fd_info
->handleType
==
2971 VK_EXTERNAL_MEMORY_HANDLE_TYPE_DMA_BUF_BIT_EXT
);
2973 result
= anv_bo_cache_import(device
, &device
->bo_cache
, fd_info
->fd
,
2974 bo_flags
| ANV_BO_EXTERNAL
, &mem
->bo
);
2975 if (result
!= VK_SUCCESS
)
2978 VkDeviceSize aligned_alloc_size
=
2979 align_u64(pAllocateInfo
->allocationSize
, 4096);
2981 /* For security purposes, we reject importing the bo if it's smaller
2982 * than the requested allocation size. This prevents a malicious client
2983 * from passing a buffer to a trusted client, lying about the size, and
2984 * telling the trusted client to try and texture from an image that goes
2985 * out-of-bounds. This sort of thing could lead to GPU hangs or worse
2986 * in the trusted client. The trusted client can protect itself against
2987 * this sort of attack but only if it can trust the buffer size.
2989 if (mem
->bo
->size
< aligned_alloc_size
) {
2990 result
= vk_errorf(device
->instance
, device
,
2991 VK_ERROR_INVALID_EXTERNAL_HANDLE
,
2992 "aligned allocationSize too large for "
2993 "VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD_BIT: "
2994 "%"PRIu64
"B > %"PRIu64
"B",
2995 aligned_alloc_size
, mem
->bo
->size
);
2996 anv_bo_cache_release(device
, &device
->bo_cache
, mem
->bo
);
3000 /* From the Vulkan spec:
3002 * "Importing memory from a file descriptor transfers ownership of
3003 * the file descriptor from the application to the Vulkan
3004 * implementation. The application must not perform any operations on
3005 * the file descriptor after a successful import."
3007 * If the import fails, we leave the file descriptor open.
3013 const VkImportMemoryHostPointerInfoEXT
*host_ptr_info
=
3014 vk_find_struct_const(pAllocateInfo
->pNext
,
3015 IMPORT_MEMORY_HOST_POINTER_INFO_EXT
);
3016 if (host_ptr_info
&& host_ptr_info
->handleType
) {
3017 if (host_ptr_info
->handleType
==
3018 VK_EXTERNAL_MEMORY_HANDLE_TYPE_HOST_MAPPED_FOREIGN_MEMORY_BIT_EXT
) {
3019 result
= vk_error(VK_ERROR_INVALID_EXTERNAL_HANDLE
);
3023 assert(host_ptr_info
->handleType
==
3024 VK_EXTERNAL_MEMORY_HANDLE_TYPE_HOST_ALLOCATION_BIT_EXT
);
3026 result
= anv_bo_cache_import_host_ptr(
3027 device
, &device
->bo_cache
, host_ptr_info
->pHostPointer
,
3028 pAllocateInfo
->allocationSize
, bo_flags
, &mem
->bo
);
3030 if (result
!= VK_SUCCESS
)
3033 mem
->host_ptr
= host_ptr_info
->pHostPointer
;
3037 /* Regular allocate (not importing memory). */
3039 if (export_info
&& export_info
->handleTypes
)
3040 bo_flags
|= ANV_BO_EXTERNAL
;
3042 result
= anv_bo_cache_alloc(device
, &device
->bo_cache
,
3043 pAllocateInfo
->allocationSize
, bo_flags
,
3045 if (result
!= VK_SUCCESS
)
3048 const VkMemoryDedicatedAllocateInfo
*dedicated_info
=
3049 vk_find_struct_const(pAllocateInfo
->pNext
, MEMORY_DEDICATED_ALLOCATE_INFO
);
3050 if (dedicated_info
&& dedicated_info
->image
!= VK_NULL_HANDLE
) {
3051 ANV_FROM_HANDLE(anv_image
, image
, dedicated_info
->image
);
3053 /* Some legacy (non-modifiers) consumers need the tiling to be set on
3054 * the BO. In this case, we have a dedicated allocation.
3056 if (image
->needs_set_tiling
) {
3057 const uint32_t i915_tiling
=
3058 isl_tiling_to_i915_tiling(image
->planes
[0].surface
.isl
.tiling
);
3059 int ret
= anv_gem_set_tiling(device
, mem
->bo
->gem_handle
,
3060 image
->planes
[0].surface
.isl
.row_pitch_B
,
3063 anv_bo_cache_release(device
, &device
->bo_cache
, mem
->bo
);
3064 return vk_errorf(device
->instance
, NULL
,
3065 VK_ERROR_OUT_OF_DEVICE_MEMORY
,
3066 "failed to set BO tiling: %m");
3072 pthread_mutex_lock(&device
->mutex
);
3073 list_addtail(&mem
->link
, &device
->memory_objects
);
3074 pthread_mutex_unlock(&device
->mutex
);
3076 *pMem
= anv_device_memory_to_handle(mem
);
3078 p_atomic_add(&pdevice
->memory
.heaps
[mem
->type
->heapIndex
].used
,
3084 vk_free2(&device
->alloc
, pAllocator
, mem
);
3089 VkResult
anv_GetMemoryFdKHR(
3091 const VkMemoryGetFdInfoKHR
* pGetFdInfo
,
3094 ANV_FROM_HANDLE(anv_device
, dev
, device_h
);
3095 ANV_FROM_HANDLE(anv_device_memory
, mem
, pGetFdInfo
->memory
);
3097 assert(pGetFdInfo
->sType
== VK_STRUCTURE_TYPE_MEMORY_GET_FD_INFO_KHR
);
3099 assert(pGetFdInfo
->handleType
== VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD_BIT
||
3100 pGetFdInfo
->handleType
== VK_EXTERNAL_MEMORY_HANDLE_TYPE_DMA_BUF_BIT_EXT
);
3102 return anv_bo_cache_export(dev
, &dev
->bo_cache
, mem
->bo
, pFd
);
3105 VkResult
anv_GetMemoryFdPropertiesKHR(
3107 VkExternalMemoryHandleTypeFlagBits handleType
,
3109 VkMemoryFdPropertiesKHR
* pMemoryFdProperties
)
3111 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3112 struct anv_physical_device
*pdevice
= &device
->instance
->physicalDevice
;
3114 switch (handleType
) {
3115 case VK_EXTERNAL_MEMORY_HANDLE_TYPE_DMA_BUF_BIT_EXT
:
3116 /* dma-buf can be imported as any memory type */
3117 pMemoryFdProperties
->memoryTypeBits
=
3118 (1 << pdevice
->memory
.type_count
) - 1;
3122 /* The valid usage section for this function says:
3124 * "handleType must not be one of the handle types defined as
3127 * So opaque handle types fall into the default "unsupported" case.
3129 return vk_error(VK_ERROR_INVALID_EXTERNAL_HANDLE
);
3133 VkResult
anv_GetMemoryHostPointerPropertiesEXT(
3135 VkExternalMemoryHandleTypeFlagBits handleType
,
3136 const void* pHostPointer
,
3137 VkMemoryHostPointerPropertiesEXT
* pMemoryHostPointerProperties
)
3139 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3141 assert(pMemoryHostPointerProperties
->sType
==
3142 VK_STRUCTURE_TYPE_MEMORY_HOST_POINTER_PROPERTIES_EXT
);
3144 switch (handleType
) {
3145 case VK_EXTERNAL_MEMORY_HANDLE_TYPE_HOST_ALLOCATION_BIT_EXT
: {
3146 struct anv_physical_device
*pdevice
= &device
->instance
->physicalDevice
;
3148 /* Host memory can be imported as any memory type. */
3149 pMemoryHostPointerProperties
->memoryTypeBits
=
3150 (1ull << pdevice
->memory
.type_count
) - 1;
3155 return VK_ERROR_INVALID_EXTERNAL_HANDLE
;
3159 void anv_FreeMemory(
3161 VkDeviceMemory _mem
,
3162 const VkAllocationCallbacks
* pAllocator
)
3164 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3165 ANV_FROM_HANDLE(anv_device_memory
, mem
, _mem
);
3166 struct anv_physical_device
*pdevice
= &device
->instance
->physicalDevice
;
3171 pthread_mutex_lock(&device
->mutex
);
3172 list_del(&mem
->link
);
3173 pthread_mutex_unlock(&device
->mutex
);
3176 anv_UnmapMemory(_device
, _mem
);
3178 p_atomic_add(&pdevice
->memory
.heaps
[mem
->type
->heapIndex
].used
,
3181 anv_bo_cache_release(device
, &device
->bo_cache
, mem
->bo
);
3183 #if defined(ANDROID) && ANDROID_API_LEVEL >= 26
3185 AHardwareBuffer_release(mem
->ahw
);
3188 vk_free2(&device
->alloc
, pAllocator
, mem
);
3191 VkResult
anv_MapMemory(
3193 VkDeviceMemory _memory
,
3194 VkDeviceSize offset
,
3196 VkMemoryMapFlags flags
,
3199 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3200 ANV_FROM_HANDLE(anv_device_memory
, mem
, _memory
);
3207 if (mem
->host_ptr
) {
3208 *ppData
= mem
->host_ptr
+ offset
;
3212 if (size
== VK_WHOLE_SIZE
)
3213 size
= mem
->bo
->size
- offset
;
3215 /* From the Vulkan spec version 1.0.32 docs for MapMemory:
3217 * * If size is not equal to VK_WHOLE_SIZE, size must be greater than 0
3218 * assert(size != 0);
3219 * * If size is not equal to VK_WHOLE_SIZE, size must be less than or
3220 * equal to the size of the memory minus offset
3223 assert(offset
+ size
<= mem
->bo
->size
);
3225 /* FIXME: Is this supposed to be thread safe? Since vkUnmapMemory() only
3226 * takes a VkDeviceMemory pointer, it seems like only one map of the memory
3227 * at a time is valid. We could just mmap up front and return an offset
3228 * pointer here, but that may exhaust virtual memory on 32 bit
3231 uint32_t gem_flags
= 0;
3233 if (!device
->info
.has_llc
&&
3234 (mem
->type
->propertyFlags
& VK_MEMORY_PROPERTY_HOST_COHERENT_BIT
))
3235 gem_flags
|= I915_MMAP_WC
;
3237 /* GEM will fail to map if the offset isn't 4k-aligned. Round down. */
3238 uint64_t map_offset
= offset
& ~4095ull;
3239 assert(offset
>= map_offset
);
3240 uint64_t map_size
= (offset
+ size
) - map_offset
;
3242 /* Let's map whole pages */
3243 map_size
= align_u64(map_size
, 4096);
3245 void *map
= anv_gem_mmap(device
, mem
->bo
->gem_handle
,
3246 map_offset
, map_size
, gem_flags
);
3247 if (map
== MAP_FAILED
)
3248 return vk_error(VK_ERROR_MEMORY_MAP_FAILED
);
3251 mem
->map_size
= map_size
;
3253 *ppData
= mem
->map
+ (offset
- map_offset
);
3258 void anv_UnmapMemory(
3260 VkDeviceMemory _memory
)
3262 ANV_FROM_HANDLE(anv_device_memory
, mem
, _memory
);
3264 if (mem
== NULL
|| mem
->host_ptr
)
3267 anv_gem_munmap(mem
->map
, mem
->map_size
);
3274 clflush_mapped_ranges(struct anv_device
*device
,
3276 const VkMappedMemoryRange
*ranges
)
3278 for (uint32_t i
= 0; i
< count
; i
++) {
3279 ANV_FROM_HANDLE(anv_device_memory
, mem
, ranges
[i
].memory
);
3280 if (ranges
[i
].offset
>= mem
->map_size
)
3283 gen_clflush_range(mem
->map
+ ranges
[i
].offset
,
3284 MIN2(ranges
[i
].size
, mem
->map_size
- ranges
[i
].offset
));
3288 VkResult
anv_FlushMappedMemoryRanges(
3290 uint32_t memoryRangeCount
,
3291 const VkMappedMemoryRange
* pMemoryRanges
)
3293 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3295 if (device
->info
.has_llc
)
3298 /* Make sure the writes we're flushing have landed. */
3299 __builtin_ia32_mfence();
3301 clflush_mapped_ranges(device
, memoryRangeCount
, pMemoryRanges
);
3306 VkResult
anv_InvalidateMappedMemoryRanges(
3308 uint32_t memoryRangeCount
,
3309 const VkMappedMemoryRange
* pMemoryRanges
)
3311 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3313 if (device
->info
.has_llc
)
3316 clflush_mapped_ranges(device
, memoryRangeCount
, pMemoryRanges
);
3318 /* Make sure no reads get moved up above the invalidate. */
3319 __builtin_ia32_mfence();
3324 void anv_GetBufferMemoryRequirements(
3327 VkMemoryRequirements
* pMemoryRequirements
)
3329 ANV_FROM_HANDLE(anv_buffer
, buffer
, _buffer
);
3330 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3331 struct anv_physical_device
*pdevice
= &device
->instance
->physicalDevice
;
3333 /* The Vulkan spec (git aaed022) says:
3335 * memoryTypeBits is a bitfield and contains one bit set for every
3336 * supported memory type for the resource. The bit `1<<i` is set if and
3337 * only if the memory type `i` in the VkPhysicalDeviceMemoryProperties
3338 * structure for the physical device is supported.
3340 uint32_t memory_types
= 0;
3341 for (uint32_t i
= 0; i
< pdevice
->memory
.type_count
; i
++) {
3342 uint32_t valid_usage
= pdevice
->memory
.types
[i
].valid_buffer_usage
;
3343 if ((valid_usage
& buffer
->usage
) == buffer
->usage
)
3344 memory_types
|= (1u << i
);
3347 /* Base alignment requirement of a cache line */
3348 uint32_t alignment
= 16;
3350 /* We need an alignment of 32 for pushing UBOs */
3351 if (buffer
->usage
& VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT
)
3352 alignment
= MAX2(alignment
, 32);
3354 pMemoryRequirements
->size
= buffer
->size
;
3355 pMemoryRequirements
->alignment
= alignment
;
3357 /* Storage and Uniform buffers should have their size aligned to
3358 * 32-bits to avoid boundary checks when last DWord is not complete.
3359 * This would ensure that not internal padding would be needed for
3362 if (device
->robust_buffer_access
&&
3363 (buffer
->usage
& VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT
||
3364 buffer
->usage
& VK_BUFFER_USAGE_STORAGE_BUFFER_BIT
))
3365 pMemoryRequirements
->size
= align_u64(buffer
->size
, 4);
3367 pMemoryRequirements
->memoryTypeBits
= memory_types
;
3370 void anv_GetBufferMemoryRequirements2(
3372 const VkBufferMemoryRequirementsInfo2
* pInfo
,
3373 VkMemoryRequirements2
* pMemoryRequirements
)
3375 anv_GetBufferMemoryRequirements(_device
, pInfo
->buffer
,
3376 &pMemoryRequirements
->memoryRequirements
);
3378 vk_foreach_struct(ext
, pMemoryRequirements
->pNext
) {
3379 switch (ext
->sType
) {
3380 case VK_STRUCTURE_TYPE_MEMORY_DEDICATED_REQUIREMENTS
: {
3381 VkMemoryDedicatedRequirements
*requirements
= (void *)ext
;
3382 requirements
->prefersDedicatedAllocation
= false;
3383 requirements
->requiresDedicatedAllocation
= false;
3388 anv_debug_ignored_stype(ext
->sType
);
3394 void anv_GetImageMemoryRequirements(
3397 VkMemoryRequirements
* pMemoryRequirements
)
3399 ANV_FROM_HANDLE(anv_image
, image
, _image
);
3400 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3401 struct anv_physical_device
*pdevice
= &device
->instance
->physicalDevice
;
3403 /* The Vulkan spec (git aaed022) says:
3405 * memoryTypeBits is a bitfield and contains one bit set for every
3406 * supported memory type for the resource. The bit `1<<i` is set if and
3407 * only if the memory type `i` in the VkPhysicalDeviceMemoryProperties
3408 * structure for the physical device is supported.
3410 * All types are currently supported for images.
3412 uint32_t memory_types
= (1ull << pdevice
->memory
.type_count
) - 1;
3414 /* We must have image allocated or imported at this point. According to the
3415 * specification, external images must have been bound to memory before
3416 * calling GetImageMemoryRequirements.
3418 assert(image
->size
> 0);
3420 pMemoryRequirements
->size
= image
->size
;
3421 pMemoryRequirements
->alignment
= image
->alignment
;
3422 pMemoryRequirements
->memoryTypeBits
= memory_types
;
3425 void anv_GetImageMemoryRequirements2(
3427 const VkImageMemoryRequirementsInfo2
* pInfo
,
3428 VkMemoryRequirements2
* pMemoryRequirements
)
3430 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3431 ANV_FROM_HANDLE(anv_image
, image
, pInfo
->image
);
3433 anv_GetImageMemoryRequirements(_device
, pInfo
->image
,
3434 &pMemoryRequirements
->memoryRequirements
);
3436 vk_foreach_struct_const(ext
, pInfo
->pNext
) {
3437 switch (ext
->sType
) {
3438 case VK_STRUCTURE_TYPE_IMAGE_PLANE_MEMORY_REQUIREMENTS_INFO
: {
3439 struct anv_physical_device
*pdevice
= &device
->instance
->physicalDevice
;
3440 const VkImagePlaneMemoryRequirementsInfo
*plane_reqs
=
3441 (const VkImagePlaneMemoryRequirementsInfo
*) ext
;
3442 uint32_t plane
= anv_image_aspect_to_plane(image
->aspects
,
3443 plane_reqs
->planeAspect
);
3445 assert(image
->planes
[plane
].offset
== 0);
3447 /* The Vulkan spec (git aaed022) says:
3449 * memoryTypeBits is a bitfield and contains one bit set for every
3450 * supported memory type for the resource. The bit `1<<i` is set
3451 * if and only if the memory type `i` in the
3452 * VkPhysicalDeviceMemoryProperties structure for the physical
3453 * device is supported.
3455 * All types are currently supported for images.
3457 pMemoryRequirements
->memoryRequirements
.memoryTypeBits
=
3458 (1ull << pdevice
->memory
.type_count
) - 1;
3460 /* We must have image allocated or imported at this point. According to the
3461 * specification, external images must have been bound to memory before
3462 * calling GetImageMemoryRequirements.
3464 assert(image
->planes
[plane
].size
> 0);
3466 pMemoryRequirements
->memoryRequirements
.size
= image
->planes
[plane
].size
;
3467 pMemoryRequirements
->memoryRequirements
.alignment
=
3468 image
->planes
[plane
].alignment
;
3473 anv_debug_ignored_stype(ext
->sType
);
3478 vk_foreach_struct(ext
, pMemoryRequirements
->pNext
) {
3479 switch (ext
->sType
) {
3480 case VK_STRUCTURE_TYPE_MEMORY_DEDICATED_REQUIREMENTS
: {
3481 VkMemoryDedicatedRequirements
*requirements
= (void *)ext
;
3482 if (image
->needs_set_tiling
|| image
->external_format
) {
3483 /* If we need to set the tiling for external consumers, we need a
3484 * dedicated allocation.
3486 * See also anv_AllocateMemory.
3488 requirements
->prefersDedicatedAllocation
= true;
3489 requirements
->requiresDedicatedAllocation
= true;
3491 requirements
->prefersDedicatedAllocation
= false;
3492 requirements
->requiresDedicatedAllocation
= false;
3498 anv_debug_ignored_stype(ext
->sType
);
3504 void anv_GetImageSparseMemoryRequirements(
3507 uint32_t* pSparseMemoryRequirementCount
,
3508 VkSparseImageMemoryRequirements
* pSparseMemoryRequirements
)
3510 *pSparseMemoryRequirementCount
= 0;
3513 void anv_GetImageSparseMemoryRequirements2(
3515 const VkImageSparseMemoryRequirementsInfo2
* pInfo
,
3516 uint32_t* pSparseMemoryRequirementCount
,
3517 VkSparseImageMemoryRequirements2
* pSparseMemoryRequirements
)
3519 *pSparseMemoryRequirementCount
= 0;
3522 void anv_GetDeviceMemoryCommitment(
3524 VkDeviceMemory memory
,
3525 VkDeviceSize
* pCommittedMemoryInBytes
)
3527 *pCommittedMemoryInBytes
= 0;
3531 anv_bind_buffer_memory(const VkBindBufferMemoryInfo
*pBindInfo
)
3533 ANV_FROM_HANDLE(anv_device_memory
, mem
, pBindInfo
->memory
);
3534 ANV_FROM_HANDLE(anv_buffer
, buffer
, pBindInfo
->buffer
);
3536 assert(pBindInfo
->sType
== VK_STRUCTURE_TYPE_BIND_BUFFER_MEMORY_INFO
);
3539 assert((buffer
->usage
& mem
->type
->valid_buffer_usage
) == buffer
->usage
);
3540 buffer
->address
= (struct anv_address
) {
3542 .offset
= pBindInfo
->memoryOffset
,
3545 buffer
->address
= ANV_NULL_ADDRESS
;
3549 VkResult
anv_BindBufferMemory(
3552 VkDeviceMemory memory
,
3553 VkDeviceSize memoryOffset
)
3555 anv_bind_buffer_memory(
3556 &(VkBindBufferMemoryInfo
) {
3557 .sType
= VK_STRUCTURE_TYPE_BIND_BUFFER_MEMORY_INFO
,
3560 .memoryOffset
= memoryOffset
,
3566 VkResult
anv_BindBufferMemory2(
3568 uint32_t bindInfoCount
,
3569 const VkBindBufferMemoryInfo
* pBindInfos
)
3571 for (uint32_t i
= 0; i
< bindInfoCount
; i
++)
3572 anv_bind_buffer_memory(&pBindInfos
[i
]);
3577 VkResult
anv_QueueBindSparse(
3579 uint32_t bindInfoCount
,
3580 const VkBindSparseInfo
* pBindInfo
,
3583 ANV_FROM_HANDLE(anv_queue
, queue
, _queue
);
3584 if (anv_device_is_lost(queue
->device
))
3585 return VK_ERROR_DEVICE_LOST
;
3587 return vk_error(VK_ERROR_FEATURE_NOT_PRESENT
);
3592 VkResult
anv_CreateEvent(
3594 const VkEventCreateInfo
* pCreateInfo
,
3595 const VkAllocationCallbacks
* pAllocator
,
3598 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3599 struct anv_state state
;
3600 struct anv_event
*event
;
3602 assert(pCreateInfo
->sType
== VK_STRUCTURE_TYPE_EVENT_CREATE_INFO
);
3604 state
= anv_state_pool_alloc(&device
->dynamic_state_pool
,
3607 event
->state
= state
;
3608 event
->semaphore
= VK_EVENT_RESET
;
3610 if (!device
->info
.has_llc
) {
3611 /* Make sure the writes we're flushing have landed. */
3612 __builtin_ia32_mfence();
3613 __builtin_ia32_clflush(event
);
3616 *pEvent
= anv_event_to_handle(event
);
3621 void anv_DestroyEvent(
3624 const VkAllocationCallbacks
* pAllocator
)
3626 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3627 ANV_FROM_HANDLE(anv_event
, event
, _event
);
3632 anv_state_pool_free(&device
->dynamic_state_pool
, event
->state
);
3635 VkResult
anv_GetEventStatus(
3639 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3640 ANV_FROM_HANDLE(anv_event
, event
, _event
);
3642 if (anv_device_is_lost(device
))
3643 return VK_ERROR_DEVICE_LOST
;
3645 if (!device
->info
.has_llc
) {
3646 /* Invalidate read cache before reading event written by GPU. */
3647 __builtin_ia32_clflush(event
);
3648 __builtin_ia32_mfence();
3652 return event
->semaphore
;
3655 VkResult
anv_SetEvent(
3659 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3660 ANV_FROM_HANDLE(anv_event
, event
, _event
);
3662 event
->semaphore
= VK_EVENT_SET
;
3664 if (!device
->info
.has_llc
) {
3665 /* Make sure the writes we're flushing have landed. */
3666 __builtin_ia32_mfence();
3667 __builtin_ia32_clflush(event
);
3673 VkResult
anv_ResetEvent(
3677 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3678 ANV_FROM_HANDLE(anv_event
, event
, _event
);
3680 event
->semaphore
= VK_EVENT_RESET
;
3682 if (!device
->info
.has_llc
) {
3683 /* Make sure the writes we're flushing have landed. */
3684 __builtin_ia32_mfence();
3685 __builtin_ia32_clflush(event
);
3693 VkResult
anv_CreateBuffer(
3695 const VkBufferCreateInfo
* pCreateInfo
,
3696 const VkAllocationCallbacks
* pAllocator
,
3699 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3700 struct anv_buffer
*buffer
;
3702 assert(pCreateInfo
->sType
== VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO
);
3704 buffer
= vk_alloc2(&device
->alloc
, pAllocator
, sizeof(*buffer
), 8,
3705 VK_SYSTEM_ALLOCATION_SCOPE_OBJECT
);
3707 return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY
);
3709 buffer
->size
= pCreateInfo
->size
;
3710 buffer
->usage
= pCreateInfo
->usage
;
3711 buffer
->address
= ANV_NULL_ADDRESS
;
3713 *pBuffer
= anv_buffer_to_handle(buffer
);
3718 void anv_DestroyBuffer(
3721 const VkAllocationCallbacks
* pAllocator
)
3723 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3724 ANV_FROM_HANDLE(anv_buffer
, buffer
, _buffer
);
3729 vk_free2(&device
->alloc
, pAllocator
, buffer
);
3732 VkDeviceAddress
anv_GetBufferDeviceAddressEXT(
3734 const VkBufferDeviceAddressInfoEXT
* pInfo
)
3736 ANV_FROM_HANDLE(anv_buffer
, buffer
, pInfo
->buffer
);
3738 assert(buffer
->address
.bo
->flags
& EXEC_OBJECT_PINNED
);
3740 return anv_address_physical(buffer
->address
);
3744 anv_fill_buffer_surface_state(struct anv_device
*device
, struct anv_state state
,
3745 enum isl_format format
,
3746 struct anv_address address
,
3747 uint32_t range
, uint32_t stride
)
3749 isl_buffer_fill_state(&device
->isl_dev
, state
.map
,
3750 .address
= anv_address_physical(address
),
3751 .mocs
= device
->default_mocs
,
3754 .swizzle
= ISL_SWIZZLE_IDENTITY
,
3755 .stride_B
= stride
);
3758 void anv_DestroySampler(
3761 const VkAllocationCallbacks
* pAllocator
)
3763 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3764 ANV_FROM_HANDLE(anv_sampler
, sampler
, _sampler
);
3769 if (sampler
->bindless_state
.map
) {
3770 anv_state_pool_free(&device
->dynamic_state_pool
,
3771 sampler
->bindless_state
);
3774 vk_free2(&device
->alloc
, pAllocator
, sampler
);
3777 VkResult
anv_CreateFramebuffer(
3779 const VkFramebufferCreateInfo
* pCreateInfo
,
3780 const VkAllocationCallbacks
* pAllocator
,
3781 VkFramebuffer
* pFramebuffer
)
3783 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3784 struct anv_framebuffer
*framebuffer
;
3786 assert(pCreateInfo
->sType
== VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO
);
3788 size_t size
= sizeof(*framebuffer
);
3790 /* VK_KHR_imageless_framebuffer extension says:
3792 * If flags includes VK_FRAMEBUFFER_CREATE_IMAGELESS_BIT_KHR,
3793 * parameter pAttachments is ignored.
3795 if (!(pCreateInfo
->flags
& VK_FRAMEBUFFER_CREATE_IMAGELESS_BIT_KHR
)) {
3796 size
+= sizeof(struct anv_image_view
*) * pCreateInfo
->attachmentCount
;
3797 framebuffer
= vk_alloc2(&device
->alloc
, pAllocator
, size
, 8,
3798 VK_SYSTEM_ALLOCATION_SCOPE_OBJECT
);
3799 if (framebuffer
== NULL
)
3800 return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY
);
3802 for (uint32_t i
= 0; i
< pCreateInfo
->attachmentCount
; i
++) {
3803 ANV_FROM_HANDLE(anv_image_view
, iview
, pCreateInfo
->pAttachments
[i
]);
3804 framebuffer
->attachments
[i
] = iview
;
3806 framebuffer
->attachment_count
= pCreateInfo
->attachmentCount
;
3808 assert(device
->enabled_extensions
.KHR_imageless_framebuffer
);
3809 framebuffer
= vk_alloc2(&device
->alloc
, pAllocator
, size
, 8,
3810 VK_SYSTEM_ALLOCATION_SCOPE_OBJECT
);
3811 if (framebuffer
== NULL
)
3812 return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY
);
3814 framebuffer
->attachment_count
= 0;
3817 framebuffer
->width
= pCreateInfo
->width
;
3818 framebuffer
->height
= pCreateInfo
->height
;
3819 framebuffer
->layers
= pCreateInfo
->layers
;
3821 *pFramebuffer
= anv_framebuffer_to_handle(framebuffer
);
3826 void anv_DestroyFramebuffer(
3829 const VkAllocationCallbacks
* pAllocator
)
3831 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3832 ANV_FROM_HANDLE(anv_framebuffer
, fb
, _fb
);
3837 vk_free2(&device
->alloc
, pAllocator
, fb
);
3840 static const VkTimeDomainEXT anv_time_domains
[] = {
3841 VK_TIME_DOMAIN_DEVICE_EXT
,
3842 VK_TIME_DOMAIN_CLOCK_MONOTONIC_EXT
,
3843 VK_TIME_DOMAIN_CLOCK_MONOTONIC_RAW_EXT
,
3846 VkResult
anv_GetPhysicalDeviceCalibrateableTimeDomainsEXT(
3847 VkPhysicalDevice physicalDevice
,
3848 uint32_t *pTimeDomainCount
,
3849 VkTimeDomainEXT
*pTimeDomains
)
3852 VK_OUTARRAY_MAKE(out
, pTimeDomains
, pTimeDomainCount
);
3854 for (d
= 0; d
< ARRAY_SIZE(anv_time_domains
); d
++) {
3855 vk_outarray_append(&out
, i
) {
3856 *i
= anv_time_domains
[d
];
3860 return vk_outarray_status(&out
);
3864 anv_clock_gettime(clockid_t clock_id
)
3866 struct timespec current
;
3869 ret
= clock_gettime(clock_id
, ¤t
);
3870 if (ret
< 0 && clock_id
== CLOCK_MONOTONIC_RAW
)
3871 ret
= clock_gettime(CLOCK_MONOTONIC
, ¤t
);
3875 return (uint64_t) current
.tv_sec
* 1000000000ULL + current
.tv_nsec
;
3878 #define TIMESTAMP 0x2358
3880 VkResult
anv_GetCalibratedTimestampsEXT(
3882 uint32_t timestampCount
,
3883 const VkCalibratedTimestampInfoEXT
*pTimestampInfos
,
3884 uint64_t *pTimestamps
,
3885 uint64_t *pMaxDeviation
)
3887 ANV_FROM_HANDLE(anv_device
, device
, _device
);
3888 uint64_t timestamp_frequency
= device
->info
.timestamp_frequency
;
3891 uint64_t begin
, end
;
3892 uint64_t max_clock_period
= 0;
3894 begin
= anv_clock_gettime(CLOCK_MONOTONIC_RAW
);
3896 for (d
= 0; d
< timestampCount
; d
++) {
3897 switch (pTimestampInfos
[d
].timeDomain
) {
3898 case VK_TIME_DOMAIN_DEVICE_EXT
:
3899 ret
= anv_gem_reg_read(device
, TIMESTAMP
| 1,
3903 return anv_device_set_lost(device
, "Failed to read the TIMESTAMP "
3906 uint64_t device_period
= DIV_ROUND_UP(1000000000, timestamp_frequency
);
3907 max_clock_period
= MAX2(max_clock_period
, device_period
);
3909 case VK_TIME_DOMAIN_CLOCK_MONOTONIC_EXT
:
3910 pTimestamps
[d
] = anv_clock_gettime(CLOCK_MONOTONIC
);
3911 max_clock_period
= MAX2(max_clock_period
, 1);
3914 case VK_TIME_DOMAIN_CLOCK_MONOTONIC_RAW_EXT
:
3915 pTimestamps
[d
] = begin
;
3923 end
= anv_clock_gettime(CLOCK_MONOTONIC_RAW
);
3926 * The maximum deviation is the sum of the interval over which we
3927 * perform the sampling and the maximum period of any sampled
3928 * clock. That's because the maximum skew between any two sampled
3929 * clock edges is when the sampled clock with the largest period is
3930 * sampled at the end of that period but right at the beginning of the
3931 * sampling interval and some other clock is sampled right at the
3932 * begining of its sampling period and right at the end of the
3933 * sampling interval. Let's assume the GPU has the longest clock
3934 * period and that the application is sampling GPU and monotonic:
3937 * w x y z 0 1 2 3 4 5 6 7 8 9 a b c d e f
3938 * Raw -_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-
3942 * GPU -----_____-----_____-----_____-----_____
3945 * x y z 0 1 2 3 4 5 6 7 8 9 a b c
3946 * Monotonic -_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-
3948 * Interval <----------------->
3949 * Deviation <-------------------------->
3953 * m = read(monotonic) 2
3956 * We round the sample interval up by one tick to cover sampling error
3957 * in the interval clock
3960 uint64_t sample_interval
= end
- begin
+ 1;
3962 *pMaxDeviation
= sample_interval
+ max_clock_period
;
3967 /* vk_icd.h does not declare this function, so we declare it here to
3968 * suppress Wmissing-prototypes.
3970 PUBLIC VKAPI_ATTR VkResult VKAPI_CALL
3971 vk_icdNegotiateLoaderICDInterfaceVersion(uint32_t* pSupportedVersion
);
3973 PUBLIC VKAPI_ATTR VkResult VKAPI_CALL
3974 vk_icdNegotiateLoaderICDInterfaceVersion(uint32_t* pSupportedVersion
)
3976 /* For the full details on loader interface versioning, see
3977 * <https://github.com/KhronosGroup/Vulkan-LoaderAndValidationLayers/blob/master/loader/LoaderAndLayerInterface.md>.
3978 * What follows is a condensed summary, to help you navigate the large and
3979 * confusing official doc.
3981 * - Loader interface v0 is incompatible with later versions. We don't
3984 * - In loader interface v1:
3985 * - The first ICD entrypoint called by the loader is
3986 * vk_icdGetInstanceProcAddr(). The ICD must statically expose this
3988 * - The ICD must statically expose no other Vulkan symbol unless it is
3989 * linked with -Bsymbolic.
3990 * - Each dispatchable Vulkan handle created by the ICD must be
3991 * a pointer to a struct whose first member is VK_LOADER_DATA. The
3992 * ICD must initialize VK_LOADER_DATA.loadMagic to ICD_LOADER_MAGIC.
3993 * - The loader implements vkCreate{PLATFORM}SurfaceKHR() and
3994 * vkDestroySurfaceKHR(). The ICD must be capable of working with
3995 * such loader-managed surfaces.
3997 * - Loader interface v2 differs from v1 in:
3998 * - The first ICD entrypoint called by the loader is
3999 * vk_icdNegotiateLoaderICDInterfaceVersion(). The ICD must
4000 * statically expose this entrypoint.
4002 * - Loader interface v3 differs from v2 in:
4003 * - The ICD must implement vkCreate{PLATFORM}SurfaceKHR(),
4004 * vkDestroySurfaceKHR(), and other API which uses VKSurfaceKHR,
4005 * because the loader no longer does so.
4007 *pSupportedVersion
= MIN2(*pSupportedVersion
, 3u);