RepoMirrors/ffmpeg
1
0
Fork 0
mirror of https://git.ffmpeg.org/ffmpeg.git synced 2024-12-28 10:22:10 +00:00
Code Issues Projects Releases Wiki Activity
e79686be96
ffmpeg/libavutil/hwcontext_vulkan.c
Philip Langdale 378fb40282 avutil/hwcontext_vulkan: disable multiplane when deriving from cuda 
Today, cuda is not able to import multiplane images, and cuda requires
images to be imported whether you trying to import to cuda or export
from cuda (in the later case, the image is imported and then copied
into on the cuda side). So any interop between cuda and vulkan requires
that multiplane be disabled.

The existing option for this is not sufficient, because when deriving
devices it is not possible to specify any options.

And, it is necessary to derive the Vulkan device, because any pipeline
that involves uploading from cuda to vulkan and then back to cuda must
use the same cuda context on both sides, and the only way to propagate
the cuda context all the way through is to derive the device at each
stage.

ie:

-vf hwupload=derive_device=vulkan,<filters>,hwupload=derive_device=cuda
2023-06-03 16:29:38 -07:00

3714 lines
144 KiB
C
Raw Blame History

/*
* Copyright (c) Lynne
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#define VK_NO_PROTOTYPES
#define VK_ENABLE_BETA_EXTENSIONS
#ifdef _WIN32
#include <windows.h> /* Included to prevent conflicts with CreateSemaphore */
#include <versionhelpers.h>
#include "compat/w32dlfcn.h"
#else
#include <dlfcn.h>
#endif
#include "thread.h"
#include <unistd.h>
#include "config.h"
#include "pixdesc.h"
#include "avstring.h"
#include "imgutils.h"
#include "hwcontext.h"
#include "avassert.h"
#include "hwcontext_internal.h"
#include "hwcontext_vulkan.h"
#include "vulkan.h"
#include "vulkan_loader.h"
#if CONFIG_VAAPI
#include "hwcontext_vaapi.h"
#endif
#if CONFIG_LIBDRM
#if CONFIG_VAAPI
#include <va/va_drmcommon.h>
#endif
#include <sys/sysmacros.h>
#include <sys/stat.h>
#include <xf86drm.h>
#include <drm_fourcc.h>
#include "hwcontext_drm.h"
#endif
#if CONFIG_CUDA
#include "hwcontext_cuda_internal.h"
#include "cuda_check.h"
#define CHECK_CU(x) FF_CUDA_CHECK_DL(cuda_cu, cu, x)
#endif
typedef struct VulkanQueueCtx {
VkFence fence;
VkQueue queue;
int was_synchronous;
int qf;
int qidx;
/* Buffer dependencies */
AVBufferRef **buf_deps;
int nb_buf_deps;
unsigned int buf_deps_alloc_size;
} VulkanQueueCtx;
typedef struct VulkanDevicePriv {
/* Vulkan library and loader functions */
void *libvulkan;
FFVulkanContext vkctx;
FFVkQueueFamilyCtx compute_qf;
FFVkQueueFamilyCtx transfer_qf;
/* Properties */
VkPhysicalDeviceProperties2 props;
VkPhysicalDeviceMemoryProperties mprops;
VkPhysicalDeviceExternalMemoryHostPropertiesEXT hprops;
/* Features */
VkPhysicalDeviceVulkan11Features device_features_1_1;
VkPhysicalDeviceVulkan12Features device_features_1_2;
VkPhysicalDeviceVulkan13Features device_features_1_3;
VkPhysicalDeviceDescriptorBufferFeaturesEXT desc_buf_features;
VkPhysicalDeviceShaderAtomicFloatFeaturesEXT atomic_float_features;
/* Queues */
pthread_mutex_t **qf_mutex;
uint32_t nb_tot_qfs;
uint32_t img_qfs[5];
uint32_t nb_img_qfs;
/* Debug callback */
VkDebugUtilsMessengerEXT debug_ctx;
/* Settings */
int use_linear_images;
/* Option to allocate all image planes in a single allocation */
int contiguous_planes;
/* Disable multiplane images */
int disable_multiplane;
/* Nvidia */
int dev_is_nvidia;
} VulkanDevicePriv;
typedef struct VulkanFramesPriv {
/* Image conversions */
FFVkExecPool compute_exec;
/* Image transfers */
FFVkExecPool upload_exec;
FFVkExecPool download_exec;
/* Modifier info list to free at uninit */
VkImageDrmFormatModifierListCreateInfoEXT *modifier_info;
} VulkanFramesPriv;
typedef struct AVVkFrameInternal {
pthread_mutex_t update_mutex;
#if CONFIG_CUDA
/* Importing external memory into cuda is really expensive so we keep the
* memory imported all the time */
AVBufferRef *cuda_fc_ref; /* Need to keep it around for uninit */
CUexternalMemory ext_mem[AV_NUM_DATA_POINTERS];
CUmipmappedArray cu_mma[AV_NUM_DATA_POINTERS];
CUarray cu_array[AV_NUM_DATA_POINTERS];
CUexternalSemaphore cu_sem[AV_NUM_DATA_POINTERS];
#ifdef _WIN32
HANDLE ext_mem_handle[AV_NUM_DATA_POINTERS];
HANDLE ext_sem_handle[AV_NUM_DATA_POINTERS];
#endif
#endif
} AVVkFrameInternal;
#define ASPECT_2PLANE (VK_IMAGE_ASPECT_PLANE_0_BIT | VK_IMAGE_ASPECT_PLANE_1_BIT)
#define ASPECT_3PLANE (VK_IMAGE_ASPECT_PLANE_0_BIT | VK_IMAGE_ASPECT_PLANE_1_BIT | VK_IMAGE_ASPECT_PLANE_2_BIT)
static const struct FFVkFormatEntry {
VkFormat vkf;
enum AVPixelFormat pixfmt;
VkImageAspectFlags aspect;
int vk_planes;
int nb_images;
int nb_images_fallback;
const VkFormat fallback[5];
} vk_formats_list[] = {
/* Gray formats */
{ VK_FORMAT_R8_UNORM, AV_PIX_FMT_GRAY8, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R8_UNORM } },
{ VK_FORMAT_R16_UNORM, AV_PIX_FMT_GRAY16, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_R32_SFLOAT, AV_PIX_FMT_GRAYF32, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R32_SFLOAT } },
/* RGB formats */
{ VK_FORMAT_R16G16B16A16_UNORM, AV_PIX_FMT_XV36, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R16G16B16A16_UNORM } },
{ VK_FORMAT_B8G8R8A8_UNORM, AV_PIX_FMT_BGRA, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_B8G8R8A8_UNORM } },
{ VK_FORMAT_R8G8B8A8_UNORM, AV_PIX_FMT_RGBA, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R8G8B8A8_UNORM } },
{ VK_FORMAT_R8G8B8_UNORM, AV_PIX_FMT_RGB24, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R8G8B8_UNORM } },
{ VK_FORMAT_B8G8R8_UNORM, AV_PIX_FMT_BGR24, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_B8G8R8_UNORM } },
{ VK_FORMAT_R16G16B16_UNORM, AV_PIX_FMT_RGB48, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R16G16B16_UNORM } },
{ VK_FORMAT_R16G16B16A16_UNORM, AV_PIX_FMT_RGBA64, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R16G16B16A16_UNORM } },
{ VK_FORMAT_R5G6B5_UNORM_PACK16, AV_PIX_FMT_RGB565, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R5G6B5_UNORM_PACK16 } },
{ VK_FORMAT_B5G6R5_UNORM_PACK16, AV_PIX_FMT_BGR565, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_B5G6R5_UNORM_PACK16 } },
{ VK_FORMAT_B8G8R8A8_UNORM, AV_PIX_FMT_BGR0, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_B8G8R8A8_UNORM } },
{ VK_FORMAT_R8G8B8A8_UNORM, AV_PIX_FMT_RGB0, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R8G8B8A8_UNORM } },
{ VK_FORMAT_A2R10G10B10_UNORM_PACK32, AV_PIX_FMT_X2RGB10, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_A2R10G10B10_UNORM_PACK32 } },
/* Planar RGB */
{ VK_FORMAT_R8_UNORM, AV_PIX_FMT_GBRAP, VK_IMAGE_ASPECT_COLOR_BIT, 1, 4, 4, { VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM } },
{ VK_FORMAT_R16_UNORM, AV_PIX_FMT_GBRAP16, VK_IMAGE_ASPECT_COLOR_BIT, 1, 4, 4, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_R32_SFLOAT, AV_PIX_FMT_GBRPF32, VK_IMAGE_ASPECT_COLOR_BIT, 1, 3, 3, { VK_FORMAT_R32_SFLOAT, VK_FORMAT_R32_SFLOAT, VK_FORMAT_R32_SFLOAT } },
{ VK_FORMAT_R32_SFLOAT, AV_PIX_FMT_GBRAPF32, VK_IMAGE_ASPECT_COLOR_BIT, 1, 4, 4, { VK_FORMAT_R32_SFLOAT, VK_FORMAT_R32_SFLOAT, VK_FORMAT_R32_SFLOAT, VK_FORMAT_R32_SFLOAT } },
/* Two-plane 420 YUV at 8, 10, 12 and 16 bits */
{ VK_FORMAT_G8_B8R8_2PLANE_420_UNORM, AV_PIX_FMT_NV12, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R8_UNORM, VK_FORMAT_R8G8_UNORM } },
{ VK_FORMAT_G10X6_B10X6R10X6_2PLANE_420_UNORM_3PACK16, AV_PIX_FMT_P010, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
{ VK_FORMAT_G12X4_B12X4R12X4_2PLANE_420_UNORM_3PACK16, AV_PIX_FMT_P012, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
{ VK_FORMAT_G16_B16R16_2PLANE_420_UNORM, AV_PIX_FMT_P016, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
/* Two-plane 422 YUV at 8, 10 and 16 bits */
{ VK_FORMAT_G8_B8R8_2PLANE_422_UNORM, AV_PIX_FMT_NV16, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R8_UNORM, VK_FORMAT_R8G8_UNORM } },
{ VK_FORMAT_G10X6_B10X6R10X6_2PLANE_422_UNORM_3PACK16, AV_PIX_FMT_P210, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
{ VK_FORMAT_G12X4_B12X4R12X4_2PLANE_422_UNORM_3PACK16, AV_PIX_FMT_P212, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
{ VK_FORMAT_G16_B16R16_2PLANE_422_UNORM, AV_PIX_FMT_P216, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
/* Two-plane 444 YUV at 8, 10 and 16 bits */
{ VK_FORMAT_G8_B8R8_2PLANE_444_UNORM, AV_PIX_FMT_NV24, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R8_UNORM, VK_FORMAT_R8G8_UNORM } },
{ VK_FORMAT_G10X6_B10X6R10X6_2PLANE_444_UNORM_3PACK16, AV_PIX_FMT_P410, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
{ VK_FORMAT_G12X4_B12X4R12X4_2PLANE_444_UNORM_3PACK16, AV_PIX_FMT_P412, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
{ VK_FORMAT_G16_B16R16_2PLANE_444_UNORM, AV_PIX_FMT_P416, ASPECT_2PLANE, 2, 1, 2, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16G16_UNORM } },
/* Three-plane 420, 422, 444 at 8, 10, 12 and 16 bits */
{ VK_FORMAT_G8_B8_R8_3PLANE_420_UNORM, AV_PIX_FMT_YUV420P, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_420_UNORM, AV_PIX_FMT_YUV420P10, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_420_UNORM, AV_PIX_FMT_YUV420P12, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_420_UNORM, AV_PIX_FMT_YUV420P16, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_G8_B8_R8_3PLANE_422_UNORM, AV_PIX_FMT_YUV422P, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_422_UNORM, AV_PIX_FMT_YUV422P10, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_422_UNORM, AV_PIX_FMT_YUV422P12, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_422_UNORM, AV_PIX_FMT_YUV422P16, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_G8_B8_R8_3PLANE_444_UNORM, AV_PIX_FMT_YUV444P, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM, VK_FORMAT_R8_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_444_UNORM, AV_PIX_FMT_YUV444P10, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_444_UNORM, AV_PIX_FMT_YUV444P12, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
{ VK_FORMAT_G16_B16_R16_3PLANE_444_UNORM, AV_PIX_FMT_YUV444P16, ASPECT_3PLANE, 3, 1, 3, { VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UNORM } },
/* Single plane 422 at 8, 10 and 12 bits */
{ VK_FORMAT_G8B8G8R8_422_UNORM, AV_PIX_FMT_YUYV422, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R8G8B8A8_UNORM } },
{ VK_FORMAT_B8G8R8G8_422_UNORM, AV_PIX_FMT_UYVY422, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R8G8B8A8_UNORM } },
{ VK_FORMAT_G10X6B10X6G10X6R10X6_422_UNORM_4PACK16, AV_PIX_FMT_Y210, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R16G16B16A16_UNORM } },
{ VK_FORMAT_G12X4B12X4G12X4R12X4_422_UNORM_4PACK16, AV_PIX_FMT_Y212, VK_IMAGE_ASPECT_COLOR_BIT, 1, 1, 1, { VK_FORMAT_R16G16B16A16_UNORM } },
};
static const int nb_vk_formats_list = FF_ARRAY_ELEMS(vk_formats_list);
const VkFormat *av_vkfmt_from_pixfmt(enum AVPixelFormat p)
{
for (int i = 0; i < nb_vk_formats_list; i++)
if (vk_formats_list[i].pixfmt == p)
return vk_formats_list[i].fallback;
return NULL;
}
static const struct FFVkFormatEntry *vk_find_format_entry(enum AVPixelFormat p)
{
for (int i = 0; i < nb_vk_formats_list; i++)
if (vk_formats_list[i].pixfmt == p)
return &vk_formats_list[i];
return NULL;
}
/* Malitia pura, Khronos */
#define FN_MAP_TO(dst_t, dst_name, src_t, src_name) \
static av_unused dst_t map_ ##src_name## _to_ ##dst_name(src_t src) \
{ \
dst_t dst = 0x0; \
MAP_TO(VK_FORMAT_FEATURE_2_SAMPLED_IMAGE_BIT, \
VK_IMAGE_USAGE_SAMPLED_BIT); \
MAP_TO(VK_FORMAT_FEATURE_2_TRANSFER_SRC_BIT, \
VK_IMAGE_USAGE_TRANSFER_SRC_BIT); \
MAP_TO(VK_FORMAT_FEATURE_2_TRANSFER_DST_BIT, \
VK_IMAGE_USAGE_TRANSFER_DST_BIT); \
MAP_TO(VK_FORMAT_FEATURE_2_STORAGE_IMAGE_BIT, \
VK_IMAGE_USAGE_STORAGE_BIT); \
MAP_TO(VK_FORMAT_FEATURE_2_COLOR_ATTACHMENT_BIT, \
VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT); \
MAP_TO(VK_FORMAT_FEATURE_2_VIDEO_DECODE_OUTPUT_BIT_KHR, \
VK_IMAGE_USAGE_VIDEO_DECODE_DST_BIT_KHR); \
MAP_TO(VK_FORMAT_FEATURE_2_VIDEO_DECODE_DPB_BIT_KHR, \
VK_IMAGE_USAGE_VIDEO_DECODE_DPB_BIT_KHR); \
MAP_TO(VK_FORMAT_FEATURE_2_VIDEO_ENCODE_DPB_BIT_KHR, \
VK_IMAGE_USAGE_VIDEO_ENCODE_DPB_BIT_KHR); \
MAP_TO(VK_FORMAT_FEATURE_2_VIDEO_ENCODE_INPUT_BIT_KHR, \
VK_IMAGE_USAGE_VIDEO_ENCODE_SRC_BIT_KHR); \
return dst; \
}
#define MAP_TO(flag1, flag2) if (src & flag2) dst |= flag1;
FN_MAP_TO(VkFormatFeatureFlagBits2, feats, VkImageUsageFlags, usage)
#undef MAP_TO
#define MAP_TO(flag1, flag2) if (src & flag1) dst |= flag2;
FN_MAP_TO(VkImageUsageFlags, usage, VkFormatFeatureFlagBits2, feats)
#undef MAP_TO
#undef FN_MAP_TO
static int vkfmt_from_pixfmt2(AVHWDeviceContext *dev_ctx, enum AVPixelFormat p,
VkImageTiling tiling,
VkFormat fmts[AV_NUM_DATA_POINTERS],
int *nb_images, VkImageAspectFlags *aspect,
VkImageUsageFlags *supported_usage, int disable_multiplane)
{
AVVulkanDeviceContext *hwctx = dev_ctx->hwctx;
VulkanDevicePriv *priv = dev_ctx->internal->priv;
FFVulkanFunctions *vk = &priv->vkctx.vkfn;
const VkFormatFeatureFlagBits2 basic_flags = VK_FORMAT_FEATURE_2_SAMPLED_IMAGE_BIT |
VK_FORMAT_FEATURE_2_TRANSFER_SRC_BIT |
VK_FORMAT_FEATURE_2_TRANSFER_DST_BIT;
for (int i = 0; i < nb_vk_formats_list; i++) {
if (vk_formats_list[i].pixfmt == p) {
VkFormatProperties2 prop = {
.sType = VK_STRUCTURE_TYPE_FORMAT_PROPERTIES_2,
};
VkFormatFeatureFlagBits2 feats_primary, feats_secondary;
int basics_primary = 0, basics_secondary = 0;
vk->GetPhysicalDeviceFormatProperties2(hwctx->phys_dev,
vk_formats_list[i].vkf,
&prop);
feats_primary = tiling == VK_IMAGE_TILING_LINEAR ?
prop.formatProperties.linearTilingFeatures :
prop.formatProperties.optimalTilingFeatures;
basics_primary = (feats_primary & basic_flags) == basic_flags;
if (vk_formats_list[i].vkf != vk_formats_list[i].fallback[0]) {
vk->GetPhysicalDeviceFormatProperties2(hwctx->phys_dev,
vk_formats_list[i].fallback[0],
&prop);
feats_secondary = tiling == VK_IMAGE_TILING_LINEAR ?
prop.formatProperties.linearTilingFeatures :
prop.formatProperties.optimalTilingFeatures;
basics_secondary = (feats_secondary & basic_flags) == basic_flags;
} else {
basics_secondary = basics_primary;
}
if (basics_primary && !(disable_multiplane && vk_formats_list[i].vk_planes > 1)) {
if (fmts)
fmts[0] = vk_formats_list[i].vkf;
if (nb_images)
*nb_images = 1;
if (aspect)
*aspect = vk_formats_list[i].aspect;
if (supported_usage)
*supported_usage = map_feats_to_usage(feats_primary);
return 0;
} else if (basics_secondary) {
if (fmts) {
for (int j = 0; j < vk_formats_list[i].nb_images_fallback; j++)
fmts[j] = vk_formats_list[i].fallback[j];
}
if (nb_images)
*nb_images = vk_formats_list[i].nb_images_fallback;
if (aspect)
*aspect = vk_formats_list[i].aspect;
if (supported_usage)
*supported_usage = map_feats_to_usage(feats_secondary);
return 0;
} else {
return AVERROR(ENOTSUP);
}
}
}
return AVERROR(EINVAL);
}
static int load_libvulkan(AVHWDeviceContext *ctx)
{
AVVulkanDeviceContext *hwctx = ctx->hwctx;
VulkanDevicePriv *p = ctx->internal->priv;
static const char *lib_names[] = {
#if defined(_WIN32)
"vulkan-1.dll",
#elif defined(__APPLE__)
"libvulkan.dylib",
"libvulkan.1.dylib",
"libMoltenVK.dylib",
#else
"libvulkan.so.1",
"libvulkan.so",
#endif
};
for (int i = 0; i < FF_ARRAY_ELEMS(lib_names); i++) {
p->libvulkan = dlopen(lib_names[i], RTLD_NOW | RTLD_LOCAL);
if (p->libvulkan)
break;
}
if (!p->libvulkan) {
av_log(ctx, AV_LOG_ERROR, "Unable to open the libvulkan library!\n");
return AVERROR_UNKNOWN;
}
hwctx->get_proc_addr = (PFN_vkGetInstanceProcAddr)dlsym(p->libvulkan, "vkGetInstanceProcAddr");
return 0;
}
typedef struct VulkanOptExtension {
const char *name;
FFVulkanExtensions flag;
} VulkanOptExtension;
static const VulkanOptExtension optional_instance_exts[] = {
/* Pointless, here avoid zero-sized structs */
{ VK_KHR_PORTABILITY_ENUMERATION_EXTENSION_NAME, FF_VK_EXT_NO_FLAG },
};
static const VulkanOptExtension optional_device_exts[] = {
/* Misc or required by other extensions */
{ VK_KHR_PORTABILITY_SUBSET_EXTENSION_NAME, FF_VK_EXT_NO_FLAG },
{ VK_KHR_PUSH_DESCRIPTOR_EXTENSION_NAME, FF_VK_EXT_NO_FLAG },
{ VK_KHR_SAMPLER_YCBCR_CONVERSION_EXTENSION_NAME, FF_VK_EXT_NO_FLAG },
{ VK_EXT_DESCRIPTOR_BUFFER_EXTENSION_NAME, FF_VK_EXT_DESCRIPTOR_BUFFER, },
{ VK_EXT_PHYSICAL_DEVICE_DRM_EXTENSION_NAME, FF_VK_EXT_DEVICE_DRM },
{ VK_EXT_SHADER_ATOMIC_FLOAT_EXTENSION_NAME, FF_VK_EXT_ATOMIC_FLOAT },
/* Imports/exports */
{ VK_KHR_EXTERNAL_MEMORY_FD_EXTENSION_NAME, FF_VK_EXT_EXTERNAL_FD_MEMORY },
{ VK_EXT_EXTERNAL_MEMORY_DMA_BUF_EXTENSION_NAME, FF_VK_EXT_EXTERNAL_DMABUF_MEMORY },
{ VK_EXT_IMAGE_DRM_FORMAT_MODIFIER_EXTENSION_NAME, FF_VK_EXT_DRM_MODIFIER_FLAGS },
{ VK_KHR_EXTERNAL_SEMAPHORE_FD_EXTENSION_NAME, FF_VK_EXT_EXTERNAL_FD_SEM },
{ VK_EXT_EXTERNAL_MEMORY_HOST_EXTENSION_NAME, FF_VK_EXT_EXTERNAL_HOST_MEMORY },
#ifdef _WIN32
{ VK_KHR_EXTERNAL_MEMORY_WIN32_EXTENSION_NAME, FF_VK_EXT_EXTERNAL_WIN32_MEMORY },
{ VK_KHR_EXTERNAL_SEMAPHORE_WIN32_EXTENSION_NAME, FF_VK_EXT_EXTERNAL_WIN32_SEM },
#endif
/* Video encoding/decoding */
{ VK_KHR_VIDEO_QUEUE_EXTENSION_NAME, FF_VK_EXT_VIDEO_QUEUE },
{ VK_KHR_VIDEO_DECODE_QUEUE_EXTENSION_NAME, FF_VK_EXT_VIDEO_DECODE_QUEUE },
{ VK_KHR_VIDEO_DECODE_H264_EXTENSION_NAME, FF_VK_EXT_VIDEO_DECODE_H264 },
{ VK_KHR_VIDEO_DECODE_H265_EXTENSION_NAME, FF_VK_EXT_VIDEO_DECODE_H265 },
{ "VK_MESA_video_decode_av1", FF_VK_EXT_VIDEO_DECODE_AV1 },
};
static VkBool32 VKAPI_CALL vk_dbg_callback(VkDebugUtilsMessageSeverityFlagBitsEXT severity,
VkDebugUtilsMessageTypeFlagsEXT messageType,
const VkDebugUtilsMessengerCallbackDataEXT *data,
void *priv)
{
int l;
AVHWDeviceContext *ctx = priv;
switch (severity) {
case VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT: l = AV_LOG_VERBOSE; break;
case VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT: l = AV_LOG_INFO; break;
case VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT: l = AV_LOG_WARNING; break;
case VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT: l = AV_LOG_ERROR; break;
default: l = AV_LOG_DEBUG; break;
}
av_log(ctx, l, "%s\n", data->pMessage);
for (int i = 0; i < data->cmdBufLabelCount; i++)
av_log(ctx, l, "\t%i: %s\n", i, data->pCmdBufLabels[i].pLabelName);
return 0;
}
#define ADD_VAL_TO_LIST(list, count, val) \
do { \
list = av_realloc_array(list, sizeof(*list), ++count); \
if (!list) { \
err = AVERROR(ENOMEM); \
goto fail; \
} \
list[count - 1] = av_strdup(val); \
if (!list[count - 1]) { \
err = AVERROR(ENOMEM); \
goto fail; \
} \
} while(0)
#define RELEASE_PROPS(props, count) \
if (props) { \
for (int i = 0; i < count; i++) \
av_free((void *)((props)[i])); \
av_free((void *)props); \
}
static int check_extensions(AVHWDeviceContext *ctx, int dev, AVDictionary *opts,
const char * const **dst, uint32_t *num, int debug)
{
const char *tstr;
const char **extension_names = NULL;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
int err = 0, found, extensions_found = 0;
const char *mod;
int optional_exts_num;
uint32_t sup_ext_count;
char *user_exts_str = NULL;
AVDictionaryEntry *user_exts;
VkExtensionProperties *sup_ext;
const VulkanOptExtension *optional_exts;
if (!dev) {
mod = "instance";
optional_exts = optional_instance_exts;
optional_exts_num = FF_ARRAY_ELEMS(optional_instance_exts);
user_exts = av_dict_get(opts, "instance_extensions", NULL, 0);
if (user_exts) {
user_exts_str = av_strdup(user_exts->value);
if (!user_exts_str) {
err = AVERROR(ENOMEM);
goto fail;
}
}
vk->EnumerateInstanceExtensionProperties(NULL, &sup_ext_count, NULL);
sup_ext = av_malloc_array(sup_ext_count, sizeof(VkExtensionProperties));
if (!sup_ext)
return AVERROR(ENOMEM);
vk->EnumerateInstanceExtensionProperties(NULL, &sup_ext_count, sup_ext);
} else {
mod = "device";
optional_exts = optional_device_exts;
optional_exts_num = FF_ARRAY_ELEMS(optional_device_exts);
user_exts = av_dict_get(opts, "device_extensions", NULL, 0);
if (user_exts) {
user_exts_str = av_strdup(user_exts->value);
if (!user_exts_str) {
err = AVERROR(ENOMEM);
goto fail;
}
}
vk->EnumerateDeviceExtensionProperties(hwctx->phys_dev, NULL,
&sup_ext_count, NULL);
sup_ext = av_malloc_array(sup_ext_count, sizeof(VkExtensionProperties));
if (!sup_ext)
return AVERROR(ENOMEM);
vk->EnumerateDeviceExtensionProperties(hwctx->phys_dev, NULL,
&sup_ext_count, sup_ext);
}
for (int i = 0; i < optional_exts_num; i++) {
tstr = optional_exts[i].name;
found = 0;
for (int j = 0; j < sup_ext_count; j++) {
if (!strcmp(tstr, sup_ext[j].extensionName)) {
found = 1;
break;
}
}
if (!found)
continue;
av_log(ctx, AV_LOG_VERBOSE, "Using %s extension %s\n", mod, tstr);
p->vkctx.extensions |= optional_exts[i].flag;
ADD_VAL_TO_LIST(extension_names, extensions_found, tstr);
}
if (debug && !dev) {
tstr = VK_EXT_DEBUG_UTILS_EXTENSION_NAME;
found = 0;
for (int j = 0; j < sup_ext_count; j++) {
if (!strcmp(tstr, sup_ext[j].extensionName)) {
found = 1;
break;
}
}
if (found) {
av_log(ctx, AV_LOG_VERBOSE, "Using %s extension %s\n", mod, tstr);
ADD_VAL_TO_LIST(extension_names, extensions_found, tstr);
p->vkctx.extensions |= FF_VK_EXT_DEBUG_UTILS;
} else {
av_log(ctx, AV_LOG_ERROR, "Debug extension \"%s\" not found!\n",
tstr);
err = AVERROR(EINVAL);
goto fail;
}
}
if (user_exts_str) {
char *save, *token = av_strtok(user_exts_str, "+", &save);
while (token) {
found = 0;
for (int j = 0; j < sup_ext_count; j++) {
if (!strcmp(token, sup_ext[j].extensionName)) {
found = 1;
break;
}
}
if (found) {
av_log(ctx, AV_LOG_VERBOSE, "Using %s extension \"%s\"\n", mod, token);
ADD_VAL_TO_LIST(extension_names, extensions_found, token);
} else {
av_log(ctx, AV_LOG_WARNING, "%s extension \"%s\" not found, excluding.\n",
mod, token);
}
token = av_strtok(NULL, "+", &save);
}
}
*dst = extension_names;
*num = extensions_found;
av_free(user_exts_str);
av_free(sup_ext);
return 0;
fail:
RELEASE_PROPS(extension_names, extensions_found);
av_free(user_exts_str);
av_free(sup_ext);
return err;
}
static int check_validation_layers(AVHWDeviceContext *ctx, AVDictionary *opts,
const char * const **dst, uint32_t *num,
int *debug_mode)
{
static const char default_layer[] = { "VK_LAYER_KHRONOS_validation" };
int found = 0, err = 0;
VulkanDevicePriv *priv = ctx->internal->priv;
FFVulkanFunctions *vk = &priv->vkctx.vkfn;
uint32_t sup_layer_count;
VkLayerProperties *sup_layers;
AVDictionaryEntry *user_layers;
char *user_layers_str = NULL;
char *save, *token;
const char **enabled_layers = NULL;
uint32_t enabled_layers_count = 0;
AVDictionaryEntry *debug_opt = av_dict_get(opts, "debug", NULL, 0);
int debug = debug_opt && strtol(debug_opt->value, NULL, 10);
/* If `debug=0`, enable no layers at all. */
if (debug_opt && !debug)
return 0;
vk->EnumerateInstanceLayerProperties(&sup_layer_count, NULL);
sup_layers = av_malloc_array(sup_layer_count, sizeof(VkLayerProperties));
if (!sup_layers)
return AVERROR(ENOMEM);
vk->EnumerateInstanceLayerProperties(&sup_layer_count, sup_layers);
av_log(ctx, AV_LOG_VERBOSE, "Supported validation layers:\n");
for (int i = 0; i < sup_layer_count; i++)
av_log(ctx, AV_LOG_VERBOSE, "\t%s\n", sup_layers[i].layerName);
/* If `debug=1` is specified, enable the standard validation layer extension */
if (debug) {
*debug_mode = debug;
for (int i = 0; i < sup_layer_count; i++) {
if (!strcmp(default_layer, sup_layers[i].layerName)) {
found = 1;
av_log(ctx, AV_LOG_VERBOSE, "Default validation layer %s is enabled\n",
default_layer);
ADD_VAL_TO_LIST(enabled_layers, enabled_layers_count, default_layer);
break;
}
}
}
user_layers = av_dict_get(opts, "validation_layers", NULL, 0);
if (!user_layers)
goto end;
user_layers_str = av_strdup(user_layers->value);
if (!user_layers_str) {
err = AVERROR(ENOMEM);
goto fail;
}
token = av_strtok(user_layers_str, "+", &save);
while (token) {
found = 0;
if (!strcmp(default_layer, token)) {
if (debug) {
/* if the `debug=1`, default_layer is enabled, skip here */
token = av_strtok(NULL, "+", &save);
continue;
} else {
/* if the `debug=0`, enable debug mode to load its callback properly */
*debug_mode = debug;
}
}
for (int j = 0; j < sup_layer_count; j++) {
if (!strcmp(token, sup_layers[j].layerName)) {
found = 1;
break;
}
}
if (found) {
av_log(ctx, AV_LOG_VERBOSE, "Requested Validation Layer: %s\n", token);
ADD_VAL_TO_LIST(enabled_layers, enabled_layers_count, token);
} else {
av_log(ctx, AV_LOG_ERROR,
"Validation Layer \"%s\" not support.\n", token);
err = AVERROR(EINVAL);
goto fail;
}
token = av_strtok(NULL, "+", &save);
}
av_free(user_layers_str);
end:
av_free(sup_layers);
*dst = enabled_layers;
*num = enabled_layers_count;
return 0;
fail:
RELEASE_PROPS(enabled_layers, enabled_layers_count);
av_free(sup_layers);
av_free(user_layers_str);
return err;
}
/* Creates a VkInstance */
static int create_instance(AVHWDeviceContext *ctx, AVDictionary *opts)
{
int err = 0, debug_mode = 0;
VkResult ret;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
VkApplicationInfo application_info = {
.sType = VK_STRUCTURE_TYPE_APPLICATION_INFO,
.pApplicationName = "ffmpeg",
.applicationVersion = VK_MAKE_VERSION(LIBAVUTIL_VERSION_MAJOR,
LIBAVUTIL_VERSION_MINOR,
LIBAVUTIL_VERSION_MICRO),
.pEngineName = "libavutil",
.apiVersion = VK_API_VERSION_1_3,
.engineVersion = VK_MAKE_VERSION(LIBAVUTIL_VERSION_MAJOR,
LIBAVUTIL_VERSION_MINOR,
LIBAVUTIL_VERSION_MICRO),
};
VkValidationFeaturesEXT validation_features = {
.sType = VK_STRUCTURE_TYPE_VALIDATION_FEATURES_EXT,
};
VkInstanceCreateInfo inst_props = {
.sType = VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO,
.pApplicationInfo = &application_info,
};
if (!hwctx->get_proc_addr) {
err = load_libvulkan(ctx);
if (err < 0)
return err;
}
err = ff_vk_load_functions(ctx, vk, p->vkctx.extensions, 0, 0);
if (err < 0) {
av_log(ctx, AV_LOG_ERROR, "Unable to load instance enumeration functions!\n");
return err;
}
err = check_validation_layers(ctx, opts, &inst_props.ppEnabledLayerNames,
&inst_props.enabledLayerCount, &debug_mode);
if (err)
goto fail;
/* Check for present/missing extensions */
err = check_extensions(ctx, 0, opts, &inst_props.ppEnabledExtensionNames,
&inst_props.enabledExtensionCount, debug_mode);
hwctx->enabled_inst_extensions = inst_props.ppEnabledExtensionNames;
hwctx->nb_enabled_inst_extensions = inst_props.enabledExtensionCount;
if (err < 0)
goto fail;
if (debug_mode) {
VkValidationFeatureEnableEXT feat_list[] = {
VK_VALIDATION_FEATURE_ENABLE_GPU_ASSISTED_EXT,
VK_VALIDATION_FEATURE_ENABLE_GPU_ASSISTED_RESERVE_BINDING_SLOT_EXT,
VK_VALIDATION_FEATURE_ENABLE_SYNCHRONIZATION_VALIDATION_EXT,
};
validation_features.pEnabledValidationFeatures = feat_list;
validation_features.enabledValidationFeatureCount = FF_ARRAY_ELEMS(feat_list);
inst_props.pNext = &validation_features;
}
/* Try to create the instance */
ret = vk->CreateInstance(&inst_props, hwctx->alloc, &hwctx->inst);
/* Check for errors */
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Instance creation failure: %s\n",
ff_vk_ret2str(ret));
err = AVERROR_EXTERNAL;
goto fail;
}
err = ff_vk_load_functions(ctx, vk, p->vkctx.extensions, 1, 0);
if (err < 0) {
av_log(ctx, AV_LOG_ERROR, "Unable to load instance functions!\n");
goto fail;
}
if (debug_mode) {
VkDebugUtilsMessengerCreateInfoEXT dbg = {
.sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_MESSENGER_CREATE_INFO_EXT,
.messageSeverity = VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT,
.messageType = VK_DEBUG_UTILS_MESSAGE_TYPE_GENERAL_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_TYPE_VALIDATION_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_TYPE_PERFORMANCE_BIT_EXT,
.pfnUserCallback = vk_dbg_callback,
.pUserData = ctx,
};
vk->CreateDebugUtilsMessengerEXT(hwctx->inst, &dbg,
hwctx->alloc, &p->debug_ctx);
}
err = 0;
fail:
RELEASE_PROPS(inst_props.ppEnabledLayerNames, inst_props.enabledLayerCount);
return err;
}
typedef struct VulkanDeviceSelection {
uint8_t uuid[VK_UUID_SIZE]; /* Will use this first unless !has_uuid */
int has_uuid;
uint32_t drm_major; /* Will use this second unless !has_drm */
uint32_t drm_minor; /* Will use this second unless !has_drm */
uint32_t has_drm; /* has drm node info */
const char *name; /* Will use this third unless NULL */
uint32_t pci_device; /* Will use this fourth unless 0x0 */
uint32_t vendor_id; /* Last resort to find something deterministic */
int index; /* Finally fall back to index */
} VulkanDeviceSelection;
static const char *vk_dev_type(enum VkPhysicalDeviceType type)
{
switch (type) {
case VK_PHYSICAL_DEVICE_TYPE_INTEGRATED_GPU: return "integrated";
case VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU: return "discrete";
case VK_PHYSICAL_DEVICE_TYPE_VIRTUAL_GPU: return "virtual";
case VK_PHYSICAL_DEVICE_TYPE_CPU: return "software";
default: return "unknown";
}
}
/* Finds a device */
static int find_device(AVHWDeviceContext *ctx, VulkanDeviceSelection *select)
{
int err = 0, choice = -1;
uint32_t num;
VkResult ret;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
VkPhysicalDevice *devices = NULL;
VkPhysicalDeviceIDProperties *idp = NULL;
VkPhysicalDeviceProperties2 *prop = NULL;
VkPhysicalDeviceDrmPropertiesEXT *drm_prop = NULL;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
ret = vk->EnumeratePhysicalDevices(hwctx->inst, &num, NULL);
if (ret != VK_SUCCESS || !num) {
av_log(ctx, AV_LOG_ERROR, "No devices found: %s!\n", ff_vk_ret2str(ret));
return AVERROR(ENODEV);
}
devices = av_malloc_array(num, sizeof(VkPhysicalDevice));
if (!devices)
return AVERROR(ENOMEM);
ret = vk->EnumeratePhysicalDevices(hwctx->inst, &num, devices);
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Failed enumerating devices: %s\n",
ff_vk_ret2str(ret));
err = AVERROR(ENODEV);
goto end;
}
prop = av_calloc(num, sizeof(*prop));
if (!prop) {
err = AVERROR(ENOMEM);
goto end;
}
idp = av_calloc(num, sizeof(*idp));
if (!idp) {
err = AVERROR(ENOMEM);
goto end;
}
if (p->vkctx.extensions & FF_VK_EXT_DEVICE_DRM) {
drm_prop = av_calloc(num, sizeof(*drm_prop));
if (!drm_prop) {
err = AVERROR(ENOMEM);
goto end;
}
}
av_log(ctx, AV_LOG_VERBOSE, "GPU listing:\n");
for (int i = 0; i < num; i++) {
if (p->vkctx.extensions & FF_VK_EXT_DEVICE_DRM) {
drm_prop[i].sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DRM_PROPERTIES_EXT;
idp[i].pNext = &drm_prop[i];
}
idp[i].sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_ID_PROPERTIES;
prop[i].sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2;
prop[i].pNext = &idp[i];
vk->GetPhysicalDeviceProperties2(devices[i], &prop[i]);
av_log(ctx, AV_LOG_VERBOSE, " %d: %s (%s) (0x%x)\n", i,
prop[i].properties.deviceName,
vk_dev_type(prop[i].properties.deviceType),
prop[i].properties.deviceID);
}
if (select->has_uuid) {
for (int i = 0; i < num; i++) {
if (!strncmp(idp[i].deviceUUID, select->uuid, VK_UUID_SIZE)) {
choice = i;
goto end;
}
}
av_log(ctx, AV_LOG_ERROR, "Unable to find device by given UUID!\n");
err = AVERROR(ENODEV);
goto end;
} else if ((p->vkctx.extensions & FF_VK_EXT_DEVICE_DRM) && select->has_drm) {
for (int i = 0; i < num; i++) {
if ((select->drm_major == drm_prop[i].primaryMajor &&
select->drm_minor == drm_prop[i].primaryMinor) ||
(select->drm_major == drm_prop[i].renderMajor &&
select->drm_minor == drm_prop[i].renderMinor)) {
choice = i;
goto end;
}
}
av_log(ctx, AV_LOG_ERROR, "Unable to find device by given DRM node numbers %i:%i!\n",
select->drm_major, select->drm_minor);
err = AVERROR(ENODEV);
goto end;
} else if (select->name) {
av_log(ctx, AV_LOG_VERBOSE, "Requested device: %s\n", select->name);
for (int i = 0; i < num; i++) {
if (strstr(prop[i].properties.deviceName, select->name)) {
choice = i;
goto end;
}
}
av_log(ctx, AV_LOG_ERROR, "Unable to find device \"%s\"!\n",
select->name);
err = AVERROR(ENODEV);
goto end;
} else if (select->pci_device) {
av_log(ctx, AV_LOG_VERBOSE, "Requested device: 0x%x\n", select->pci_device);
for (int i = 0; i < num; i++) {
if (select->pci_device == prop[i].properties.deviceID) {
choice = i;
goto end;
}
}
av_log(ctx, AV_LOG_ERROR, "Unable to find device with PCI ID 0x%x!\n",
select->pci_device);
err = AVERROR(EINVAL);
goto end;
} else if (select->vendor_id) {
av_log(ctx, AV_LOG_VERBOSE, "Requested vendor: 0x%x\n", select->vendor_id);
for (int i = 0; i < num; i++) {
if (select->vendor_id == prop[i].properties.vendorID) {
choice = i;
goto end;
}
}
av_log(ctx, AV_LOG_ERROR, "Unable to find device with Vendor ID 0x%x!\n",
select->vendor_id);
err = AVERROR(ENODEV);
goto end;
} else {
if (select->index < num) {
choice = select->index;
goto end;
}
av_log(ctx, AV_LOG_ERROR, "Unable to find device with index %i!\n",
select->index);
err = AVERROR(ENODEV);
goto end;
}
end:
if (choice > -1) {
av_log(ctx, AV_LOG_VERBOSE, "Device %d selected: %s (%s) (0x%x)\n",
choice, prop[choice].properties.deviceName,
vk_dev_type(prop[choice].properties.deviceType),
prop[choice].properties.deviceID);
hwctx->phys_dev = devices[choice];
}
av_free(devices);
av_free(prop);
av_free(idp);
av_free(drm_prop);
return err;
}
/* Picks the least used qf with the fewest unneeded flags, or -1 if none found */
static inline int pick_queue_family(VkQueueFamilyProperties *qf, uint32_t num_qf,
VkQueueFlagBits flags)
{
int index = -1;
uint32_t min_score = UINT32_MAX;
for (int i = 0; i < num_qf; i++) {
const VkQueueFlagBits qflags = qf[i].queueFlags;
if (qflags & flags) {
uint32_t score = av_popcount(qflags) + qf[i].timestampValidBits;
if (score < min_score) {
index = i;
min_score = score;
}
}
}
if (index > -1)
qf[index].timestampValidBits++;
return index;
}
static int setup_queue_families(AVHWDeviceContext *ctx, VkDeviceCreateInfo *cd)
{
uint32_t num;
float *weights;
VkQueueFamilyProperties *qf = NULL;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
int graph_index, comp_index, tx_index, enc_index, dec_index;
/* First get the number of queue families */
vk->GetPhysicalDeviceQueueFamilyProperties(hwctx->phys_dev, &num, NULL);
if (!num) {
av_log(ctx, AV_LOG_ERROR, "Failed to get queues!\n");
return AVERROR_EXTERNAL;
}
/* Then allocate memory */
qf = av_malloc_array(num, sizeof(VkQueueFamilyProperties));
if (!qf)
return AVERROR(ENOMEM);
/* Finally retrieve the queue families */
vk->GetPhysicalDeviceQueueFamilyProperties(hwctx->phys_dev, &num, qf);
av_log(ctx, AV_LOG_VERBOSE, "Queue families:\n");
for (int i = 0; i < num; i++) {
av_log(ctx, AV_LOG_VERBOSE, " %i:%s%s%s%s%s%s%s (queues: %i)\n", i,
((qf[i].queueFlags) & VK_QUEUE_GRAPHICS_BIT) ? " graphics" : "",
((qf[i].queueFlags) & VK_QUEUE_COMPUTE_BIT) ? " compute" : "",
((qf[i].queueFlags) & VK_QUEUE_TRANSFER_BIT) ? " transfer" : "",
((qf[i].queueFlags) & VK_QUEUE_VIDEO_ENCODE_BIT_KHR) ? " encode" : "",
((qf[i].queueFlags) & VK_QUEUE_VIDEO_DECODE_BIT_KHR) ? " decode" : "",
((qf[i].queueFlags) & VK_QUEUE_SPARSE_BINDING_BIT) ? " sparse" : "",
((qf[i].queueFlags) & VK_QUEUE_PROTECTED_BIT) ? " protected" : "",
qf[i].queueCount);
/* We use this field to keep a score of how many times we've used that
* queue family in order to make better choices. */
qf[i].timestampValidBits = 0;
}
/* Pick each queue family to use */
graph_index = pick_queue_family(qf, num, VK_QUEUE_GRAPHICS_BIT);
comp_index = pick_queue_family(qf, num, VK_QUEUE_COMPUTE_BIT);
tx_index = pick_queue_family(qf, num, VK_QUEUE_TRANSFER_BIT);
enc_index = pick_queue_family(qf, num, VK_QUEUE_VIDEO_ENCODE_BIT_KHR);
dec_index = pick_queue_family(qf, num, VK_QUEUE_VIDEO_DECODE_BIT_KHR);
/* Signalling the transfer capabilities on a queue family is optional */
if (tx_index < 0) {
tx_index = pick_queue_family(qf, num, VK_QUEUE_COMPUTE_BIT);
if (tx_index < 0)
tx_index = pick_queue_family(qf, num, VK_QUEUE_GRAPHICS_BIT);
}
hwctx->queue_family_index = -1;
hwctx->queue_family_comp_index = -1;
hwctx->queue_family_tx_index = -1;
hwctx->queue_family_encode_index = -1;
hwctx->queue_family_decode_index = -1;
#define SETUP_QUEUE(qf_idx) \
if (qf_idx > -1) { \
int fidx = qf_idx; \
int qc = qf[fidx].queueCount; \
VkDeviceQueueCreateInfo *pc; \
\
if (fidx == graph_index) { \
hwctx->queue_family_index = fidx; \
hwctx->nb_graphics_queues = qc; \
graph_index = -1; \
} \
if (fidx == comp_index) { \
hwctx->queue_family_comp_index = fidx; \
hwctx->nb_comp_queues = qc; \
comp_index = -1; \
} \
if (fidx == tx_index) { \
hwctx->queue_family_tx_index = fidx; \
hwctx->nb_tx_queues = qc; \
tx_index = -1; \
} \
if (fidx == enc_index) { \
hwctx->queue_family_encode_index = fidx; \
hwctx->nb_encode_queues = qc; \
enc_index = -1; \
} \
if (fidx == dec_index) { \
hwctx->queue_family_decode_index = fidx; \
hwctx->nb_decode_queues = qc; \
dec_index = -1; \
} \
\
pc = av_realloc((void *)cd->pQueueCreateInfos, \
sizeof(*pc) * (cd->queueCreateInfoCount + 1)); \
if (!pc) { \
av_free(qf); \
return AVERROR(ENOMEM); \
} \
cd->pQueueCreateInfos = pc; \
pc = &pc[cd->queueCreateInfoCount]; \
\
weights = av_malloc(qc * sizeof(float)); \
if (!weights) { \
av_free(qf); \
return AVERROR(ENOMEM); \
} \
\
memset(pc, 0, sizeof(*pc)); \
pc->sType = VK_STRUCTURE_TYPE_DEVICE_QUEUE_CREATE_INFO; \
pc->queueFamilyIndex = fidx; \
pc->queueCount = qc; \
pc->pQueuePriorities = weights; \
\
for (int i = 0; i < qc; i++) \
weights[i] = 1.0f / qc; \
\
cd->queueCreateInfoCount++; \
}
SETUP_QUEUE(graph_index)
SETUP_QUEUE(comp_index)
SETUP_QUEUE(tx_index)
SETUP_QUEUE(enc_index)
SETUP_QUEUE(dec_index)
#undef SETUP_QUEUE
av_free(qf);
return 0;
}
static void vulkan_device_free(AVHWDeviceContext *ctx)
{
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
if (hwctx->act_dev)
vk->DestroyDevice(hwctx->act_dev, hwctx->alloc);
if (p->debug_ctx)
vk->DestroyDebugUtilsMessengerEXT(hwctx->inst, p->debug_ctx,
hwctx->alloc);
if (hwctx->inst)
vk->DestroyInstance(hwctx->inst, hwctx->alloc);
if (p->libvulkan)
dlclose(p->libvulkan);
for (uint32_t i = 0; i < p->nb_tot_qfs; i++) {
pthread_mutex_destroy(p->qf_mutex[i]);
av_freep(&p->qf_mutex[i]);
}
av_freep(&p->qf_mutex);
RELEASE_PROPS(hwctx->enabled_inst_extensions, hwctx->nb_enabled_inst_extensions);
RELEASE_PROPS(hwctx->enabled_dev_extensions, hwctx->nb_enabled_dev_extensions);
}
static int vulkan_device_create_internal(AVHWDeviceContext *ctx,
VulkanDeviceSelection *dev_select,
int disable_multiplane,
AVDictionary *opts, int flags)
{
int err = 0;
VkResult ret;
AVDictionaryEntry *opt_d;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
/*
* VkPhysicalDeviceVulkan12Features has a timelineSemaphore field, but
* MoltenVK doesn't implement VkPhysicalDeviceVulkan12Features yet, so we
* use VkPhysicalDeviceTimelineSemaphoreFeatures directly.
*/
VkPhysicalDeviceTimelineSemaphoreFeatures timeline_features = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TIMELINE_SEMAPHORE_FEATURES,
};
VkPhysicalDeviceShaderAtomicFloatFeaturesEXT atomic_float_features = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SHADER_ATOMIC_FLOAT_FEATURES_EXT,
.pNext = &timeline_features,
};
VkPhysicalDeviceDescriptorBufferFeaturesEXT desc_buf_features = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DESCRIPTOR_BUFFER_FEATURES_EXT,
.pNext = &atomic_float_features,
};
VkPhysicalDeviceVulkan13Features dev_features_1_3 = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VULKAN_1_3_FEATURES,
.pNext = &desc_buf_features,
};
VkPhysicalDeviceVulkan12Features dev_features_1_2 = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VULKAN_1_2_FEATURES,
.pNext = &dev_features_1_3,
};
VkPhysicalDeviceVulkan11Features dev_features_1_1 = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VULKAN_1_1_FEATURES,
.pNext = &dev_features_1_2,
};
VkPhysicalDeviceFeatures2 dev_features = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FEATURES_2,
.pNext = &dev_features_1_1,
};
VkDeviceCreateInfo dev_info = {
.sType = VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO,
};
hwctx->device_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FEATURES_2;
hwctx->device_features.pNext = &p->device_features_1_1;
p->device_features_1_1.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VULKAN_1_1_FEATURES;
p->device_features_1_1.pNext = &p->device_features_1_2;
p->device_features_1_2.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VULKAN_1_2_FEATURES;
p->device_features_1_2.pNext = &p->device_features_1_3;
p->device_features_1_3.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VULKAN_1_3_FEATURES;
p->device_features_1_3.pNext = &p->desc_buf_features;
p->desc_buf_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DESCRIPTOR_BUFFER_FEATURES_EXT;
p->desc_buf_features.pNext = &p->atomic_float_features;
p->atomic_float_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SHADER_ATOMIC_FLOAT_FEATURES_EXT;
p->atomic_float_features.pNext = NULL;
ctx->free = vulkan_device_free;
/* Create an instance if not given one */
if ((err = create_instance(ctx, opts)))
goto end;
/* Find a device (if not given one) */
if ((err = find_device(ctx, dev_select)))
goto end;
vk->GetPhysicalDeviceFeatures2(hwctx->phys_dev, &dev_features);
/* Try to keep in sync with libplacebo */
#define COPY_FEATURE(DST, NAME) (DST).features.NAME = dev_features.features.NAME;
COPY_FEATURE(hwctx->device_features, shaderImageGatherExtended)
COPY_FEATURE(hwctx->device_features, shaderStorageImageReadWithoutFormat)
COPY_FEATURE(hwctx->device_features, shaderStorageImageWriteWithoutFormat)
COPY_FEATURE(hwctx->device_features, fragmentStoresAndAtomics)
COPY_FEATURE(hwctx->device_features, vertexPipelineStoresAndAtomics)
COPY_FEATURE(hwctx->device_features, shaderInt64)
COPY_FEATURE(hwctx->device_features, shaderInt16)
COPY_FEATURE(hwctx->device_features, shaderFloat64)
#undef COPY_FEATURE
/* We require timeline semaphores */
if (!timeline_features.timelineSemaphore) {
av_log(ctx, AV_LOG_ERROR, "Device does not support timeline semaphores!\n");
err = AVERROR(ENOSYS);
goto end;
}
p->device_features_1_1.samplerYcbcrConversion = dev_features_1_1.samplerYcbcrConversion;
p->device_features_1_1.storagePushConstant16 = dev_features_1_1.storagePushConstant16;
p->device_features_1_2.timelineSemaphore = 1;
p->device_features_1_2.bufferDeviceAddress = dev_features_1_2.bufferDeviceAddress;
p->device_features_1_2.hostQueryReset = dev_features_1_2.hostQueryReset;
p->device_features_1_2.storagePushConstant8 = dev_features_1_2.storagePushConstant8;
p->device_features_1_2.shaderInt8 = dev_features_1_2.shaderInt8;
p->device_features_1_2.storageBuffer8BitAccess = dev_features_1_2.storageBuffer8BitAccess;
p->device_features_1_2.uniformAndStorageBuffer8BitAccess = dev_features_1_2.uniformAndStorageBuffer8BitAccess;
p->device_features_1_2.shaderFloat16 = dev_features_1_2.shaderFloat16;
p->device_features_1_2.shaderSharedInt64Atomics = dev_features_1_2.shaderSharedInt64Atomics;
p->device_features_1_2.vulkanMemoryModel = dev_features_1_2.vulkanMemoryModel;
p->device_features_1_2.vulkanMemoryModelDeviceScope = dev_features_1_2.vulkanMemoryModelDeviceScope;
p->device_features_1_2.hostQueryReset = dev_features_1_2.hostQueryReset;
p->device_features_1_3.dynamicRendering = dev_features_1_3.dynamicRendering;
p->device_features_1_3.maintenance4 = dev_features_1_3.maintenance4;
p->device_features_1_3.synchronization2 = dev_features_1_3.synchronization2;
p->device_features_1_3.computeFullSubgroups = dev_features_1_3.computeFullSubgroups;
p->device_features_1_3.shaderZeroInitializeWorkgroupMemory = dev_features_1_3.shaderZeroInitializeWorkgroupMemory;
p->device_features_1_3.dynamicRendering = dev_features_1_3.dynamicRendering;
p->desc_buf_features.descriptorBuffer = desc_buf_features.descriptorBuffer;
p->desc_buf_features.descriptorBufferPushDescriptors = desc_buf_features.descriptorBufferPushDescriptors;
p->atomic_float_features.shaderBufferFloat32Atomics = atomic_float_features.shaderBufferFloat32Atomics;
p->atomic_float_features.shaderBufferFloat32AtomicAdd = atomic_float_features.shaderBufferFloat32AtomicAdd;
dev_info.pNext = &hwctx->device_features;
/* Setup queue family */
if ((err = setup_queue_families(ctx, &dev_info)))
goto end;
if ((err = check_extensions(ctx, 1, opts, &dev_info.ppEnabledExtensionNames,
&dev_info.enabledExtensionCount, 0))) {
for (int i = 0; i < dev_info.queueCreateInfoCount; i++)
av_free((void *)dev_info.pQueueCreateInfos[i].pQueuePriorities);
av_free((void *)dev_info.pQueueCreateInfos);
goto end;
}
ret = vk->CreateDevice(hwctx->phys_dev, &dev_info, hwctx->alloc,
&hwctx->act_dev);
for (int i = 0; i < dev_info.queueCreateInfoCount; i++)
av_free((void *)dev_info.pQueueCreateInfos[i].pQueuePriorities);
av_free((void *)dev_info.pQueueCreateInfos);
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Device creation failure: %s\n",
ff_vk_ret2str(ret));
for (int i = 0; i < dev_info.enabledExtensionCount; i++)
av_free((void *)dev_info.ppEnabledExtensionNames[i]);
av_free((void *)dev_info.ppEnabledExtensionNames);
err = AVERROR_EXTERNAL;
goto end;
}
/* Tiled images setting, use them by default */
opt_d = av_dict_get(opts, "linear_images", NULL, 0);
if (opt_d)
p->use_linear_images = strtol(opt_d->value, NULL, 10);
/*
* The disable_multiplane argument takes precedent over the option.
*/
p->disable_multiplane = disable_multiplane;
if (!p->disable_multiplane) {
opt_d = av_dict_get(opts, "disable_multiplane", NULL, 0);
if (opt_d)
p->disable_multiplane = strtol(opt_d->value, NULL, 10);
}
hwctx->enabled_dev_extensions = dev_info.ppEnabledExtensionNames;
hwctx->nb_enabled_dev_extensions = dev_info.enabledExtensionCount;
end:
return err;
}
static void lock_queue(AVHWDeviceContext *ctx, uint32_t queue_family, uint32_t index)
{
VulkanDevicePriv *p = ctx->internal->priv;
pthread_mutex_lock(&p->qf_mutex[queue_family][index]);
}
static void unlock_queue(AVHWDeviceContext *ctx, uint32_t queue_family, uint32_t index)
{
VulkanDevicePriv *p = ctx->internal->priv;
pthread_mutex_unlock(&p->qf_mutex[queue_family][index]);
}
static int vulkan_device_init(AVHWDeviceContext *ctx)
{
int err;
uint32_t qf_num;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
VkQueueFamilyProperties *qf;
int graph_index, comp_index, tx_index, enc_index, dec_index;
/* Set device extension flags */
for (int i = 0; i < hwctx->nb_enabled_dev_extensions; i++) {
for (int j = 0; j < FF_ARRAY_ELEMS(optional_device_exts); j++) {
if (!strcmp(hwctx->enabled_dev_extensions[i],
optional_device_exts[j].name)) {
p->vkctx.extensions |= optional_device_exts[j].flag;
break;
}
}
}
err = ff_vk_load_functions(ctx, vk, p->vkctx.extensions, 1, 1);
if (err < 0) {
av_log(ctx, AV_LOG_ERROR, "Unable to load functions!\n");
return err;
}
p->props.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2;
p->props.pNext = &p->hprops;
p->hprops.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_EXTERNAL_MEMORY_HOST_PROPERTIES_EXT;
vk->GetPhysicalDeviceProperties2(hwctx->phys_dev, &p->props);
av_log(ctx, AV_LOG_VERBOSE, "Using device: %s\n",
p->props.properties.deviceName);
av_log(ctx, AV_LOG_VERBOSE, "Alignments:\n");
av_log(ctx, AV_LOG_VERBOSE, " optimalBufferCopyRowPitchAlignment: %"PRIu64"\n",
p->props.properties.limits.optimalBufferCopyRowPitchAlignment);
av_log(ctx, AV_LOG_VERBOSE, " minMemoryMapAlignment: %"SIZE_SPECIFIER"\n",
p->props.properties.limits.minMemoryMapAlignment);
av_log(ctx, AV_LOG_VERBOSE, " nonCoherentAtomSize: %"PRIu64"\n",
p->props.properties.limits.nonCoherentAtomSize);
if (p->vkctx.extensions & FF_VK_EXT_EXTERNAL_HOST_MEMORY)
av_log(ctx, AV_LOG_VERBOSE, " minImportedHostPointerAlignment: %"PRIu64"\n",
p->hprops.minImportedHostPointerAlignment);
p->dev_is_nvidia = (p->props.properties.vendorID == 0x10de);
vk->GetPhysicalDeviceQueueFamilyProperties(hwctx->phys_dev, &qf_num, NULL);
if (!qf_num) {
av_log(ctx, AV_LOG_ERROR, "Failed to get queues!\n");
return AVERROR_EXTERNAL;
}
qf = av_malloc_array(qf_num, sizeof(VkQueueFamilyProperties));
if (!qf)
return AVERROR(ENOMEM);
vk->GetPhysicalDeviceQueueFamilyProperties(hwctx->phys_dev, &qf_num, qf);
p->qf_mutex = av_calloc(qf_num, sizeof(*p->qf_mutex));
if (!p->qf_mutex)
return AVERROR(ENOMEM);
p->nb_tot_qfs = qf_num;
for (uint32_t i = 0; i < qf_num; i++) {
p->qf_mutex[i] = av_calloc(qf[i].queueCount, sizeof(**p->qf_mutex));
if (!p->qf_mutex[i])
return AVERROR(ENOMEM);
for (uint32_t j = 0; j < qf[i].queueCount; j++) {
err = pthread_mutex_init(&p->qf_mutex[i][j], NULL);
if (err != 0) {
av_log(ctx, AV_LOG_ERROR, "pthread_mutex_init failed : %s\n",
av_err2str(err));
return AVERROR(err);
}
}
}
graph_index = hwctx->queue_family_index;
comp_index = hwctx->queue_family_comp_index;
tx_index = hwctx->queue_family_tx_index;
enc_index = hwctx->queue_family_encode_index;
dec_index = hwctx->queue_family_decode_index;
#define CHECK_QUEUE(type, required, fidx, ctx_qf, qc) \
do { \
if (ctx_qf < 0 && required) { \
av_log(ctx, AV_LOG_ERROR, "%s queue family is required, but marked as missing" \
" in the context!\n", type); \
return AVERROR(EINVAL); \
} else if (fidx < 0 || ctx_qf < 0) { \
break; \
} else if (ctx_qf >= qf_num) { \
av_log(ctx, AV_LOG_ERROR, "Invalid %s family index %i (device has %i families)!\n", \
type, ctx_qf, qf_num); \
return AVERROR(EINVAL); \
} \
\
av_log(ctx, AV_LOG_VERBOSE, "Using queue family %i (queues: %i)" \
" for%s%s%s%s%s\n", \
ctx_qf, qc, \
ctx_qf == graph_index ? " graphics" : "", \
ctx_qf == comp_index ? " compute" : "", \
ctx_qf == tx_index ? " transfers" : "", \
ctx_qf == enc_index ? " encode" : "", \
ctx_qf == dec_index ? " decode" : ""); \
graph_index = (ctx_qf == graph_index) ? -1 : graph_index; \
comp_index = (ctx_qf == comp_index) ? -1 : comp_index; \
tx_index = (ctx_qf == tx_index) ? -1 : tx_index; \
enc_index = (ctx_qf == enc_index) ? -1 : enc_index; \
dec_index = (ctx_qf == dec_index) ? -1 : dec_index; \
p->img_qfs[p->nb_img_qfs++] = ctx_qf; \
} while (0)
CHECK_QUEUE("graphics", 0, graph_index, hwctx->queue_family_index, hwctx->nb_graphics_queues);
CHECK_QUEUE("upload", 1, tx_index, hwctx->queue_family_tx_index, hwctx->nb_tx_queues);
CHECK_QUEUE("compute", 1, comp_index, hwctx->queue_family_comp_index, hwctx->nb_comp_queues);
CHECK_QUEUE("encode", 0, enc_index, hwctx->queue_family_encode_index, hwctx->nb_encode_queues);
CHECK_QUEUE("decode", 0, dec_index, hwctx->queue_family_decode_index, hwctx->nb_decode_queues);
#undef CHECK_QUEUE
if (!hwctx->lock_queue)
hwctx->lock_queue = lock_queue;
if (!hwctx->unlock_queue)
hwctx->unlock_queue = unlock_queue;
/* Get device capabilities */
vk->GetPhysicalDeviceMemoryProperties(hwctx->phys_dev, &p->mprops);
p->vkctx.device = ctx;
p->vkctx.hwctx = hwctx;
ff_vk_load_props(&p->vkctx);
ff_vk_qf_init(&p->vkctx, &p->compute_qf, VK_QUEUE_COMPUTE_BIT);
ff_vk_qf_init(&p->vkctx, &p->transfer_qf, VK_QUEUE_TRANSFER_BIT);
return 0;
}
static int vulkan_device_create(AVHWDeviceContext *ctx, const char *device,
AVDictionary *opts, int flags)
{
VulkanDeviceSelection dev_select = { 0 };
if (device && device[0]) {
char *end = NULL;
dev_select.index = strtol(device, &end, 10);
if (end == device) {
dev_select.index = 0;
dev_select.name = device;
}
}
return vulkan_device_create_internal(ctx, &dev_select, 0, opts, flags);
}
static int vulkan_device_derive(AVHWDeviceContext *ctx,
AVHWDeviceContext *src_ctx,
AVDictionary *opts, int flags)
{
av_unused VulkanDeviceSelection dev_select = { 0 };
/* If there's only one device on the system, then even if its not covered
* by the following checks (e.g. non-PCIe ARM GPU), having an empty
* dev_select will mean it'll get picked. */
switch(src_ctx->type) {
#if CONFIG_VAAPI
case AV_HWDEVICE_TYPE_VAAPI: {
AVVAAPIDeviceContext *src_hwctx = src_ctx->hwctx;
const char *vendor = vaQueryVendorString(src_hwctx->display);
if (!vendor) {
av_log(ctx, AV_LOG_ERROR, "Unable to get device info from VAAPI!\n");
return AVERROR_EXTERNAL;
}
if (strstr(vendor, "AMD"))
dev_select.vendor_id = 0x1002;
return vulkan_device_create_internal(ctx, &dev_select, 0, opts, flags);
}
#endif
#if CONFIG_LIBDRM
case AV_HWDEVICE_TYPE_DRM: {
int err;
struct stat drm_node_info;
drmDevice *drm_dev_info;
AVDRMDeviceContext *src_hwctx = src_ctx->hwctx;
err = fstat(src_hwctx->fd, &drm_node_info);
if (err) {
av_log(ctx, AV_LOG_ERROR, "Unable to get node info from DRM fd: %s!\n",
av_err2str(AVERROR(errno)));
return AVERROR_EXTERNAL;
}
dev_select.drm_major = major(drm_node_info.st_dev);
dev_select.drm_minor = minor(drm_node_info.st_dev);
dev_select.has_drm = 1;
err = drmGetDevice(src_hwctx->fd, &drm_dev_info);
if (err) {
av_log(ctx, AV_LOG_ERROR, "Unable to get device info from DRM fd: %s!\n",
av_err2str(AVERROR(errno)));
return AVERROR_EXTERNAL;
}
if (drm_dev_info->bustype == DRM_BUS_PCI)
dev_select.pci_device = drm_dev_info->deviceinfo.pci->device_id;
drmFreeDevice(&drm_dev_info);
return vulkan_device_create_internal(ctx, &dev_select, 0, opts, flags);
}
#endif
#if CONFIG_CUDA
case AV_HWDEVICE_TYPE_CUDA: {
AVHWDeviceContext *cuda_cu = src_ctx;
AVCUDADeviceContext *src_hwctx = src_ctx->hwctx;
AVCUDADeviceContextInternal *cu_internal = src_hwctx->internal;
CudaFunctions *cu = cu_internal->cuda_dl;
int ret = CHECK_CU(cu->cuDeviceGetUuid((CUuuid *)&dev_select.uuid,
cu_internal->cuda_device));
if (ret < 0) {
av_log(ctx, AV_LOG_ERROR, "Unable to get UUID from CUDA!\n");
return AVERROR_EXTERNAL;
}
dev_select.has_uuid = 1;
/*
* CUDA is not able to import multiplane images, so always derive a
* Vulkan device with multiplane disabled.
*/
return vulkan_device_create_internal(ctx, &dev_select, 1, opts, flags);
}
#endif
default:
return AVERROR(ENOSYS);
}
}
static int vulkan_frames_get_constraints(AVHWDeviceContext *ctx,
const void *hwconfig,
AVHWFramesConstraints *constraints)
{
int count = 0;
VulkanDevicePriv *p = ctx->internal->priv;
for (enum AVPixelFormat i = 0; i < nb_vk_formats_list; i++) {
count += vkfmt_from_pixfmt2(ctx, vk_formats_list[i].pixfmt,
p->use_linear_images ? VK_IMAGE_TILING_LINEAR :
VK_IMAGE_TILING_OPTIMAL,
NULL, NULL, NULL, NULL, 0) >= 0;
}
#if CONFIG_CUDA
if (p->dev_is_nvidia)
count++;
#endif
constraints->valid_sw_formats = av_malloc_array(count + 1,
sizeof(enum AVPixelFormat));
if (!constraints->valid_sw_formats)
return AVERROR(ENOMEM);
count = 0;
for (enum AVPixelFormat i = 0; i < nb_vk_formats_list; i++) {
if (vkfmt_from_pixfmt2(ctx, vk_formats_list[i].pixfmt,
p->use_linear_images ? VK_IMAGE_TILING_LINEAR :
VK_IMAGE_TILING_OPTIMAL,
NULL, NULL, NULL, NULL, 0) >= 0) {
constraints->valid_sw_formats[count++] = vk_formats_list[i].pixfmt;
}
}
#if CONFIG_CUDA
if (p->dev_is_nvidia)
constraints->valid_sw_formats[count++] = AV_PIX_FMT_CUDA;
#endif
constraints->valid_sw_formats[count++] = AV_PIX_FMT_NONE;
constraints->min_width = 1;
constraints->min_height = 1;
constraints->max_width = p->props.properties.limits.maxImageDimension2D;
constraints->max_height = p->props.properties.limits.maxImageDimension2D;
constraints->valid_hw_formats = av_malloc_array(2, sizeof(enum AVPixelFormat));
if (!constraints->valid_hw_formats)
return AVERROR(ENOMEM);
constraints->valid_hw_formats[0] = AV_PIX_FMT_VULKAN;
constraints->valid_hw_formats[1] = AV_PIX_FMT_NONE;
return 0;
}
static int alloc_mem(AVHWDeviceContext *ctx, VkMemoryRequirements *req,
VkMemoryPropertyFlagBits req_flags, const void *alloc_extension,
VkMemoryPropertyFlagBits *mem_flags, VkDeviceMemory *mem)
{
VkResult ret;
int index = -1;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVVulkanDeviceContext *dev_hwctx = ctx->hwctx;
VkMemoryAllocateInfo alloc_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO,
.pNext = alloc_extension,
.allocationSize = req->size,
};
/* The vulkan spec requires memory types to be sorted in the "optimal"
* order, so the first matching type we find will be the best/fastest one */
for (int i = 0; i < p->mprops.memoryTypeCount; i++) {
const VkMemoryType *type = &p->mprops.memoryTypes[i];
/* The memory type must be supported by the requirements (bitfield) */
if (!(req->memoryTypeBits & (1 << i)))
continue;
/* The memory type flags must include our properties */
if ((type->propertyFlags & req_flags) != req_flags)
continue;
/* The memory type must be large enough */
if (req->size > p->mprops.memoryHeaps[type->heapIndex].size)
continue;
/* Found a suitable memory type */
index = i;
break;
}
if (index < 0) {
av_log(ctx, AV_LOG_ERROR, "No memory type found for flags 0x%x\n",
req_flags);
return AVERROR(EINVAL);
}
alloc_info.memoryTypeIndex = index;
ret = vk->AllocateMemory(dev_hwctx->act_dev, &alloc_info,
dev_hwctx->alloc, mem);
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Failed to allocate memory: %s\n",
ff_vk_ret2str(ret));
return AVERROR(ENOMEM);
}
*mem_flags |= p->mprops.memoryTypes[index].propertyFlags;
return 0;
}
static void vulkan_free_internal(AVVkFrame *f)
{
av_unused AVVkFrameInternal *internal = f->internal;
#if CONFIG_CUDA
if (internal->cuda_fc_ref) {
AVHWFramesContext *cuda_fc = (AVHWFramesContext *)internal->cuda_fc_ref->data;
int planes = av_pix_fmt_count_planes(cuda_fc->sw_format);
AVHWDeviceContext *cuda_cu = cuda_fc->device_ctx;
AVCUDADeviceContext *cuda_dev = cuda_cu->hwctx;
AVCUDADeviceContextInternal *cu_internal = cuda_dev->internal;
CudaFunctions *cu = cu_internal->cuda_dl;
for (int i = 0; i < planes; i++) {
if (internal->cu_sem[i])
CHECK_CU(cu->cuDestroyExternalSemaphore(internal->cu_sem[i]));
if (internal->cu_mma[i])
CHECK_CU(cu->cuMipmappedArrayDestroy(internal->cu_mma[i]));
if (internal->ext_mem[i])
CHECK_CU(cu->cuDestroyExternalMemory(internal->ext_mem[i]));
#ifdef _WIN32
if (internal->ext_sem_handle[i])
CloseHandle(internal->ext_sem_handle[i]);
if (internal->ext_mem_handle[i])
CloseHandle(internal->ext_mem_handle[i]);
#endif
}
av_buffer_unref(&internal->cuda_fc_ref);
}
#endif
pthread_mutex_destroy(&internal->update_mutex);
av_freep(&f->internal);
}
static void vulkan_frame_free(void *opaque, uint8_t *data)
{
AVVkFrame *f = (AVVkFrame *)data;
AVHWFramesContext *hwfc = opaque;
AVVulkanDeviceContext *hwctx = hwfc->device_ctx->hwctx;
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
int nb_images = ff_vk_count_images(f);
VkSemaphoreWaitInfo sem_wait = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_WAIT_INFO,
.pSemaphores = f->sem,
.pValues = f->sem_value,
.semaphoreCount = nb_images,
};
vk->WaitSemaphores(hwctx->act_dev, &sem_wait, UINT64_MAX);
vulkan_free_internal(f);
for (int i = 0; i < nb_images; i++) {
vk->DestroyImage(hwctx->act_dev, f->img[i], hwctx->alloc);
vk->FreeMemory(hwctx->act_dev, f->mem[i], hwctx->alloc);
vk->DestroySemaphore(hwctx->act_dev, f->sem[i], hwctx->alloc);
}
av_free(f);
}
static int alloc_bind_mem(AVHWFramesContext *hwfc, AVVkFrame *f,
void *alloc_pnext, size_t alloc_pnext_stride)
{
int img_cnt = 0, err;
VkResult ret;
AVHWDeviceContext *ctx = hwfc->device_ctx;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
VkBindImageMemoryInfo bind_info[AV_NUM_DATA_POINTERS] = { { 0 } };
AVVulkanDeviceContext *hwctx = ctx->hwctx;
while (f->img[img_cnt]) {
int use_ded_mem;
VkImageMemoryRequirementsInfo2 req_desc = {
.sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_REQUIREMENTS_INFO_2,
.image = f->img[img_cnt],
};
VkMemoryDedicatedAllocateInfo ded_alloc = {
.sType = VK_STRUCTURE_TYPE_MEMORY_DEDICATED_ALLOCATE_INFO,
.pNext = (void *)(((uint8_t *)alloc_pnext) + img_cnt*alloc_pnext_stride),
};
VkMemoryDedicatedRequirements ded_req = {
.sType = VK_STRUCTURE_TYPE_MEMORY_DEDICATED_REQUIREMENTS,
};
VkMemoryRequirements2 req = {
.sType = VK_STRUCTURE_TYPE_MEMORY_REQUIREMENTS_2,
.pNext = &ded_req,
};
vk->GetImageMemoryRequirements2(hwctx->act_dev, &req_desc, &req);
if (f->tiling == VK_IMAGE_TILING_LINEAR)
req.memoryRequirements.size = FFALIGN(req.memoryRequirements.size,
p->props.properties.limits.minMemoryMapAlignment);
/* In case the implementation prefers/requires dedicated allocation */
use_ded_mem = ded_req.prefersDedicatedAllocation |
ded_req.requiresDedicatedAllocation;
if (use_ded_mem)
ded_alloc.image = f->img[img_cnt];
/* Allocate memory */
if ((err = alloc_mem(ctx, &req.memoryRequirements,
f->tiling == VK_IMAGE_TILING_LINEAR ?
VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT :
VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT,
use_ded_mem ? &ded_alloc : (void *)ded_alloc.pNext,
&f->flags, &f->mem[img_cnt])))
return err;
f->size[img_cnt] = req.memoryRequirements.size;
bind_info[img_cnt].sType = VK_STRUCTURE_TYPE_BIND_IMAGE_MEMORY_INFO;
bind_info[img_cnt].image = f->img[img_cnt];
bind_info[img_cnt].memory = f->mem[img_cnt];
img_cnt++;
}
/* Bind the allocated memory to the images */
ret = vk->BindImageMemory2(hwctx->act_dev, img_cnt, bind_info);
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Failed to bind memory: %s\n",
ff_vk_ret2str(ret));
return AVERROR_EXTERNAL;
}
return 0;
}
enum PrepMode {
PREP_MODE_WRITE,
PREP_MODE_EXTERNAL_EXPORT,
PREP_MODE_EXTERNAL_IMPORT,
PREP_MODE_DECODING_DST,
PREP_MODE_DECODING_DPB,
};
static int prepare_frame(AVHWFramesContext *hwfc, FFVkExecPool *ectx,
AVVkFrame *frame, enum PrepMode pmode)
{
int err;
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
VkImageMemoryBarrier2 img_bar[AV_NUM_DATA_POINTERS];
int nb_img_bar = 0;
uint32_t dst_qf = VK_QUEUE_FAMILY_IGNORED;
VkImageLayout new_layout;
VkAccessFlags2 new_access;
VkPipelineStageFlagBits2 src_stage = VK_PIPELINE_STAGE_2_NONE;
/* This is dirty - but it works. The vulkan.c dependency system doesn't
* free non-refcounted frames, and non-refcounted hardware frames cannot
* happen anywhere outside of here. */
AVBufferRef tmp_ref = {
.data = (uint8_t *)hwfc,
};
AVFrame tmp_frame = {
.data[0] = (uint8_t *)frame,
.hw_frames_ctx = &tmp_ref,
};
VkCommandBuffer cmd_buf;
FFVkExecContext *exec = ff_vk_exec_get(ectx);
cmd_buf = exec->buf;
ff_vk_exec_start(&p->vkctx, exec);
err = ff_vk_exec_add_dep_frame(&p->vkctx, exec, &tmp_frame,
VK_PIPELINE_STAGE_2_NONE,
VK_PIPELINE_STAGE_2_ALL_COMMANDS_BIT);
if (err < 0)
return err;
switch (pmode) {
case PREP_MODE_WRITE:
new_layout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL;
new_access = VK_ACCESS_TRANSFER_WRITE_BIT;
break;
case PREP_MODE_EXTERNAL_IMPORT:
new_layout = VK_IMAGE_LAYOUT_GENERAL;
new_access = VK_ACCESS_MEMORY_READ_BIT | VK_ACCESS_MEMORY_WRITE_BIT;
break;
case PREP_MODE_EXTERNAL_EXPORT:
new_layout = VK_IMAGE_LAYOUT_GENERAL;
new_access = VK_ACCESS_MEMORY_READ_BIT | VK_ACCESS_MEMORY_WRITE_BIT;
dst_qf = VK_QUEUE_FAMILY_EXTERNAL_KHR;
src_stage = VK_PIPELINE_STAGE_2_ALL_COMMANDS_BIT;
break;
case PREP_MODE_DECODING_DST:
new_layout = VK_IMAGE_LAYOUT_VIDEO_DECODE_DST_KHR;
new_access = VK_ACCESS_TRANSFER_WRITE_BIT;
break;
case PREP_MODE_DECODING_DPB:
new_layout = VK_IMAGE_LAYOUT_VIDEO_DECODE_DPB_KHR;
new_access = VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_TRANSFER_WRITE_BIT;
break;
}
ff_vk_frame_barrier(&p->vkctx, exec, &tmp_frame, img_bar, &nb_img_bar,
src_stage,
VK_PIPELINE_STAGE_2_ALL_COMMANDS_BIT,
new_access, new_layout, dst_qf);
vk->CmdPipelineBarrier2(cmd_buf, &(VkDependencyInfo) {
.sType = VK_STRUCTURE_TYPE_DEPENDENCY_INFO,
.pImageMemoryBarriers = img_bar,
.imageMemoryBarrierCount = nb_img_bar,
});
err = ff_vk_exec_submit(&p->vkctx, exec);
if (err < 0)
return err;
/* We can do this because there are no real dependencies */
ff_vk_exec_discard_deps(&p->vkctx, exec);
return 0;
}
static inline void get_plane_wh(uint32_t *w, uint32_t *h, enum AVPixelFormat format,
int frame_w, int frame_h, int plane)
{
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(format);
/* Currently always true unless gray + alpha support is added */
if (!plane || (plane == 3) || desc->flags & AV_PIX_FMT_FLAG_RGB ||
!(desc->flags & AV_PIX_FMT_FLAG_PLANAR)) {
*w = frame_w;
*h = frame_h;
return;
}
*w = AV_CEIL_RSHIFT(frame_w, desc->log2_chroma_w);
*h = AV_CEIL_RSHIFT(frame_h, desc->log2_chroma_h);
}
static int create_frame(AVHWFramesContext *hwfc, AVVkFrame **frame,
VkImageTiling tiling, VkImageUsageFlagBits usage,
VkImageCreateFlags flags, int nb_layers,
void *create_pnext)
{
int err;
VkResult ret;
AVVulkanFramesContext *hwfc_vk = hwfc->hwctx;
AVHWDeviceContext *ctx = hwfc->device_ctx;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
VkExportSemaphoreCreateInfo ext_sem_info = {
.sType = VK_STRUCTURE_TYPE_EXPORT_SEMAPHORE_CREATE_INFO,
#ifdef _WIN32
.handleTypes = IsWindows8OrGreater()
? VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_BIT
: VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_KMT_BIT,
#else
.handleTypes = VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_FD_BIT,
#endif
};
VkSemaphoreTypeCreateInfo sem_type_info = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_TYPE_CREATE_INFO,
#ifdef _WIN32
.pNext = p->vkctx.extensions & FF_VK_EXT_EXTERNAL_WIN32_SEM ? &ext_sem_info : NULL,
#else
.pNext = p->vkctx.extensions & FF_VK_EXT_EXTERNAL_FD_SEM ? &ext_sem_info : NULL,
#endif
.semaphoreType = VK_SEMAPHORE_TYPE_TIMELINE,
.initialValue = 0,
};
VkSemaphoreCreateInfo sem_spawn = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO,
.pNext = &sem_type_info,
};
AVVkFrame *f = av_vk_frame_alloc();
if (!f) {
av_log(ctx, AV_LOG_ERROR, "Unable to allocate memory for AVVkFrame!\n");
return AVERROR(ENOMEM);
}
// TODO: check witdh and height for alignment in case of multiplanar (must be mod-2 if subsampled)
/* Create the images */
for (int i = 0; (hwfc_vk->format[i] != VK_FORMAT_UNDEFINED); i++) {
VkImageCreateInfo create_info = {
.sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO,
.pNext = create_pnext,
.imageType = VK_IMAGE_TYPE_2D,
.format = hwfc_vk->format[i],
.extent.depth = 1,
.mipLevels = 1,
.arrayLayers = nb_layers,
.flags = flags,
.tiling = tiling,
.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED,
.usage = usage,
.samples = VK_SAMPLE_COUNT_1_BIT,
.pQueueFamilyIndices = p->img_qfs,
.queueFamilyIndexCount = p->nb_img_qfs,
.sharingMode = p->nb_img_qfs > 1 ? VK_SHARING_MODE_CONCURRENT :
VK_SHARING_MODE_EXCLUSIVE,
};
get_plane_wh(&create_info.extent.width, &create_info.extent.height,
hwfc->sw_format, hwfc->width, hwfc->height, i);
ret = vk->CreateImage(hwctx->act_dev, &create_info,
hwctx->alloc, &f->img[i]);
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Image creation failure: %s\n",
ff_vk_ret2str(ret));
err = AVERROR(EINVAL);
goto fail;
}
/* Create semaphore */
ret = vk->CreateSemaphore(hwctx->act_dev, &sem_spawn,
hwctx->alloc, &f->sem[i]);
if (ret != VK_SUCCESS) {
av_log(hwctx, AV_LOG_ERROR, "Failed to create semaphore: %s\n",
ff_vk_ret2str(ret));
return AVERROR_EXTERNAL;
}
f->queue_family[i] = p->nb_img_qfs > 1 ? VK_QUEUE_FAMILY_IGNORED : p->img_qfs[0];
f->layout[i] = create_info.initialLayout;
f->access[i] = 0x0;
f->sem_value[i] = 0;
}
f->flags = 0x0;
f->tiling = tiling;
*frame = f;
return 0;
fail:
vulkan_frame_free(hwfc, (uint8_t *)f);
return err;
}
/* Checks if an export flag is enabled, and if it is ORs it with *iexp */
static void try_export_flags(AVHWFramesContext *hwfc,
VkExternalMemoryHandleTypeFlags *comp_handle_types,
VkExternalMemoryHandleTypeFlagBits *iexp,
VkExternalMemoryHandleTypeFlagBits exp)
{
VkResult ret;
AVVulkanFramesContext *hwctx = hwfc->hwctx;
AVVulkanDeviceContext *dev_hwctx = hwfc->device_ctx->hwctx;
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
const VkImageDrmFormatModifierListCreateInfoEXT *drm_mod_info =
ff_vk_find_struct(hwctx->create_pnext,
VK_STRUCTURE_TYPE_IMAGE_DRM_FORMAT_MODIFIER_LIST_CREATE_INFO_EXT);
int has_mods = hwctx->tiling == VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT && drm_mod_info;
int nb_mods;
VkExternalImageFormatProperties eprops = {
.sType = VK_STRUCTURE_TYPE_EXTERNAL_IMAGE_FORMAT_PROPERTIES_KHR,
};
VkImageFormatProperties2 props = {
.sType = VK_STRUCTURE_TYPE_IMAGE_FORMAT_PROPERTIES_2,
.pNext = &eprops,
};
VkPhysicalDeviceImageDrmFormatModifierInfoEXT phy_dev_mod_info = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_IMAGE_DRM_FORMAT_MODIFIER_INFO_EXT,
.pNext = NULL,
.pQueueFamilyIndices = p->img_qfs,
.queueFamilyIndexCount = p->nb_img_qfs,
.sharingMode = p->nb_img_qfs > 1 ? VK_SHARING_MODE_CONCURRENT :
VK_SHARING_MODE_EXCLUSIVE,
};
VkPhysicalDeviceExternalImageFormatInfo enext = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_EXTERNAL_IMAGE_FORMAT_INFO,
.handleType = exp,
.pNext = has_mods ? &phy_dev_mod_info : NULL,
};
VkPhysicalDeviceImageFormatInfo2 pinfo = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_IMAGE_FORMAT_INFO_2,
.pNext = !exp ? NULL : &enext,
.format = av_vkfmt_from_pixfmt(hwfc->sw_format)[0],
.type = VK_IMAGE_TYPE_2D,
.tiling = hwctx->tiling,
.usage = hwctx->usage,
.flags = VK_IMAGE_CREATE_ALIAS_BIT,
};
nb_mods = has_mods ? drm_mod_info->drmFormatModifierCount : 1;
for (int i = 0; i < nb_mods; i++) {
if (has_mods)
phy_dev_mod_info.drmFormatModifier = drm_mod_info->pDrmFormatModifiers[i];
ret = vk->GetPhysicalDeviceImageFormatProperties2(dev_hwctx->phys_dev,
&pinfo, &props);
if (ret == VK_SUCCESS) {
*iexp |= exp;
*comp_handle_types |= eprops.externalMemoryProperties.compatibleHandleTypes;
}
}
}
static AVBufferRef *vulkan_pool_alloc(void *opaque, size_t size)
{
int err;
AVVkFrame *f;
AVBufferRef *avbuf = NULL;
AVHWFramesContext *hwfc = opaque;
AVVulkanFramesContext *hwctx = hwfc->hwctx;
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
VulkanFramesPriv *fp = hwfc->internal->priv;
VkExternalMemoryHandleTypeFlags e = 0x0;
VkExportMemoryAllocateInfo eminfo[AV_NUM_DATA_POINTERS];
VkExternalMemoryImageCreateInfo eiinfo = {
.sType = VK_STRUCTURE_TYPE_EXTERNAL_MEMORY_IMAGE_CREATE_INFO,
.pNext = hwctx->create_pnext,
};
#ifdef _WIN32
if (p->vkctx.extensions & FF_VK_EXT_EXTERNAL_WIN32_MEMORY)
try_export_flags(hwfc, &eiinfo.handleTypes, &e, IsWindows8OrGreater()
? VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32_BIT
: VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32_KMT_BIT);
#else
if (p->vkctx.extensions & FF_VK_EXT_EXTERNAL_FD_MEMORY)
try_export_flags(hwfc, &eiinfo.handleTypes, &e,
VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD_BIT);
#endif
for (int i = 0; i < av_pix_fmt_count_planes(hwfc->sw_format); i++) {
eminfo[i].sType = VK_STRUCTURE_TYPE_EXPORT_MEMORY_ALLOCATE_INFO;
eminfo[i].pNext = hwctx->alloc_pnext[i];
eminfo[i].handleTypes = e;
}
err = create_frame(hwfc, &f, hwctx->tiling, hwctx->usage, hwctx->img_flags,
hwctx->nb_layers, eiinfo.handleTypes ? &eiinfo : NULL);
if (err)
return NULL;
err = alloc_bind_mem(hwfc, f, eminfo, sizeof(*eminfo));
if (err)
goto fail;
if ( (hwctx->usage & VK_IMAGE_USAGE_VIDEO_DECODE_DPB_BIT_KHR) &&
!(hwctx->usage & VK_IMAGE_USAGE_VIDEO_DECODE_DST_BIT_KHR))
err = prepare_frame(hwfc, &fp->compute_exec, f, PREP_MODE_DECODING_DPB);
else if (hwctx->usage & VK_IMAGE_USAGE_VIDEO_DECODE_DST_BIT_KHR)
err = prepare_frame(hwfc, &fp->compute_exec, f, PREP_MODE_DECODING_DST);
else
err = prepare_frame(hwfc, &fp->compute_exec, f, PREP_MODE_WRITE);
if (err)
goto fail;
avbuf = av_buffer_create((uint8_t *)f, sizeof(AVVkFrame),
vulkan_frame_free, hwfc, 0);
if (!avbuf)
goto fail;
return avbuf;
fail:
vulkan_frame_free(hwfc, (uint8_t *)f);
return NULL;
}
static void lock_frame(AVHWFramesContext *fc, AVVkFrame *vkf)
{
pthread_mutex_lock(&vkf->internal->update_mutex);
}
static void unlock_frame(AVHWFramesContext *fc, AVVkFrame *vkf)
{
pthread_mutex_unlock(&vkf->internal->update_mutex);
}
static void vulkan_frames_uninit(AVHWFramesContext *hwfc)
{
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
VulkanFramesPriv *fp = hwfc->internal->priv;
if (fp->modifier_info) {
if (fp->modifier_info->pDrmFormatModifiers)
av_freep(&fp->modifier_info->pDrmFormatModifiers);
av_freep(&fp->modifier_info);
}
ff_vk_exec_pool_free(&p->vkctx, &fp->compute_exec);
ff_vk_exec_pool_free(&p->vkctx, &fp->upload_exec);
ff_vk_exec_pool_free(&p->vkctx, &fp->download_exec);
}
static int vulkan_frames_init(AVHWFramesContext *hwfc)
{
int err;
AVVkFrame *f;
AVVulkanFramesContext *hwctx = hwfc->hwctx;
VulkanFramesPriv *fp = hwfc->internal->priv;
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
VkImageUsageFlagBits supported_usage;
const struct FFVkFormatEntry *fmt;
int disable_multiplane = p->disable_multiplane ||
(hwctx->flags & AV_VK_FRAME_FLAG_DISABLE_MULTIPLANE);
/* Defaults */
if (!hwctx->nb_layers)
hwctx->nb_layers = 1;
/* VK_IMAGE_TILING_OPTIMAL == 0, can't check for it really */
if (p->use_linear_images &&
(hwctx->tiling != VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT))
hwctx->tiling = VK_IMAGE_TILING_LINEAR;
fmt = vk_find_format_entry(hwfc->sw_format);
if (!fmt) {
av_log(hwfc, AV_LOG_ERROR, "Unsupported pixel format: %s!\n",
av_get_pix_fmt_name(hwfc->sw_format));
return AVERROR(EINVAL);
}
if (hwctx->format[0] != VK_FORMAT_UNDEFINED) {
if (hwctx->format[0] != fmt->vkf) {
for (int i = 0; i < fmt->nb_images_fallback; i++) {
if (hwctx->format[i] != fmt->fallback[i]) {
av_log(hwfc, AV_LOG_ERROR, "Incompatible Vulkan format given "
"for the current sw_format %s!\n",
av_get_pix_fmt_name(hwfc->sw_format));
return AVERROR(EINVAL);
}
}
}
/* Check if the sw_format itself is supported */
err = vkfmt_from_pixfmt2(hwfc->device_ctx, hwfc->sw_format,
hwctx->tiling, NULL,
NULL, NULL, &supported_usage, 0);
if (err < 0) {
av_log(hwfc, AV_LOG_ERROR, "Unsupported sw format: %s!\n",
av_get_pix_fmt_name(hwfc->sw_format));
return AVERROR(EINVAL);
}
} else {
err = vkfmt_from_pixfmt2(hwfc->device_ctx, hwfc->sw_format,
hwctx->tiling, hwctx->format, NULL,
NULL, &supported_usage,
disable_multiplane);
if (err < 0)
return err;
}
/* Image usage flags */
if (!hwctx->usage) {
hwctx->usage = supported_usage & (VK_BUFFER_USAGE_TRANSFER_DST_BIT |
VK_BUFFER_USAGE_TRANSFER_SRC_BIT |
VK_IMAGE_USAGE_STORAGE_BIT |
VK_IMAGE_USAGE_SAMPLED_BIT |
VK_IMAGE_USAGE_VIDEO_ENCODE_SRC_BIT_KHR);
}
/* Image creation flags.
* Only fill them in automatically if the image is not going to be used as
* a DPB-only image, and we have SAMPLED/STORAGE bits set. */
if (!hwctx->img_flags) {
int is_lone_dpb = (hwctx->usage & VK_IMAGE_USAGE_VIDEO_DECODE_DPB_BIT_KHR) &&
!(hwctx->usage & VK_IMAGE_USAGE_VIDEO_DECODE_DST_BIT_KHR);
int sampleable = hwctx->usage & (VK_IMAGE_USAGE_SAMPLED_BIT |
VK_IMAGE_USAGE_STORAGE_BIT);
if (sampleable && !is_lone_dpb) {
hwctx->img_flags = VK_IMAGE_CREATE_ALIAS_BIT;
if ((fmt->vk_planes > 1) && (hwctx->format[0] == fmt->vkf))
hwctx->img_flags |= VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT |
VK_IMAGE_CREATE_EXTENDED_USAGE_BIT;
}
}
if (!hwctx->lock_frame)
hwctx->lock_frame = lock_frame;
if (!hwctx->unlock_frame)
hwctx->unlock_frame = unlock_frame;
err = ff_vk_exec_pool_init(&p->vkctx, &p->compute_qf, &fp->compute_exec,
p->compute_qf.nb_queues*4, 0, 0, 0, NULL);
if (err)
return err;
err = ff_vk_exec_pool_init(&p->vkctx, &p->transfer_qf, &fp->upload_exec,
p->transfer_qf.nb_queues*4, 0, 0, 0, NULL);
if (err)
return err;
err = ff_vk_exec_pool_init(&p->vkctx, &p->transfer_qf, &fp->download_exec,
p->transfer_qf.nb_queues*4, 0, 0, 0, NULL);
if (err)
return err;
/* Test to see if allocation will fail */
err = create_frame(hwfc, &f, hwctx->tiling, hwctx->usage, hwctx->img_flags,
hwctx->nb_layers, hwctx->create_pnext);
if (err)
return err;
vulkan_frame_free(hwfc, (uint8_t *)f);
/* If user did not specify a pool, hwfc->pool will be set to the internal one
* in hwcontext.c just after this gets called */
if (!hwfc->pool) {
hwfc->internal->pool_internal = av_buffer_pool_init2(sizeof(AVVkFrame),
hwfc, vulkan_pool_alloc,
NULL);
if (!hwfc->internal->pool_internal)
return AVERROR(ENOMEM);
}
return 0;
}
static int vulkan_get_buffer(AVHWFramesContext *hwfc, AVFrame *frame)
{
frame->buf[0] = av_buffer_pool_get(hwfc->pool);
if (!frame->buf[0])
return AVERROR(ENOMEM);
frame->data[0] = frame->buf[0]->data;
frame->format = AV_PIX_FMT_VULKAN;
frame->width = hwfc->width;
frame->height = hwfc->height;
return 0;
}
static int vulkan_transfer_get_formats(AVHWFramesContext *hwfc,
enum AVHWFrameTransferDirection dir,
enum AVPixelFormat **formats)
{
enum AVPixelFormat *fmts = av_malloc_array(2, sizeof(*fmts));
if (!fmts)
return AVERROR(ENOMEM);
fmts[0] = hwfc->sw_format;
fmts[1] = AV_PIX_FMT_NONE;
*formats = fmts;
return 0;
}
#if CONFIG_LIBDRM
static void vulkan_unmap_from_drm(AVHWFramesContext *hwfc, HWMapDescriptor *hwmap)
{
AVVkFrame *f = hwmap->priv;
AVVulkanDeviceContext *hwctx = hwfc->device_ctx->hwctx;
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
const int nb_images = ff_vk_count_images(f);
VkSemaphoreWaitInfo wait_info = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_WAIT_INFO,
.flags = 0x0,
.pSemaphores = f->sem,
.pValues = f->sem_value,
.semaphoreCount = nb_images,
};
vk->WaitSemaphores(hwctx->act_dev, &wait_info, UINT64_MAX);
vulkan_free_internal(f);
for (int i = 0; i < nb_images; i++) {
vk->DestroyImage(hwctx->act_dev, f->img[i], hwctx->alloc);
vk->FreeMemory(hwctx->act_dev, f->mem[i], hwctx->alloc);
vk->DestroySemaphore(hwctx->act_dev, f->sem[i], hwctx->alloc);
}
av_free(f);
}
static const struct {
uint32_t drm_fourcc;
VkFormat vk_format;
} vulkan_drm_format_map[] = {
{ DRM_FORMAT_R8, VK_FORMAT_R8_UNORM },
{ DRM_FORMAT_R16, VK_FORMAT_R16_UNORM },
{ DRM_FORMAT_GR88, VK_FORMAT_R8G8_UNORM },
{ DRM_FORMAT_RG88, VK_FORMAT_R8G8_UNORM },
{ DRM_FORMAT_GR1616, VK_FORMAT_R16G16_UNORM },
{ DRM_FORMAT_RG1616, VK_FORMAT_R16G16_UNORM },
{ DRM_FORMAT_ARGB8888, VK_FORMAT_B8G8R8A8_UNORM },
{ DRM_FORMAT_XRGB8888, VK_FORMAT_B8G8R8A8_UNORM },
{ DRM_FORMAT_ABGR8888, VK_FORMAT_R8G8B8A8_UNORM },
{ DRM_FORMAT_XBGR8888, VK_FORMAT_R8G8B8A8_UNORM },
// All these DRM_FORMATs were added in the same libdrm commit.
#ifdef DRM_FORMAT_XYUV8888
{ DRM_FORMAT_XYUV8888, VK_FORMAT_R8G8B8A8_UNORM },
{ DRM_FORMAT_XVYU12_16161616, VK_FORMAT_R16G16B16A16_UNORM} ,
// As we had to map XV36 to a 16bit Vulkan format, reverse mapping will
// end up yielding Y416 as the DRM format, so we need to recognise it.
{ DRM_FORMAT_Y416, VK_FORMAT_R16G16B16A16_UNORM },
#endif
};
static inline VkFormat drm_to_vulkan_fmt(uint32_t drm_fourcc)
{
for (int i = 0; i < FF_ARRAY_ELEMS(vulkan_drm_format_map); i++)
if (vulkan_drm_format_map[i].drm_fourcc == drm_fourcc)
return vulkan_drm_format_map[i].vk_format;
return VK_FORMAT_UNDEFINED;
}
static int vulkan_map_from_drm_frame_desc(AVHWFramesContext *hwfc, AVVkFrame **frame,
const AVFrame *src)
{
int err = 0;
VkResult ret;
AVVkFrame *f;
int bind_counts = 0;
AVHWDeviceContext *ctx = hwfc->device_ctx;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
VulkanFramesPriv *fp = hwfc->internal->priv;
const AVDRMFrameDescriptor *desc = (AVDRMFrameDescriptor *)src->data[0];
VkBindImageMemoryInfo bind_info[AV_DRM_MAX_PLANES];
VkBindImagePlaneMemoryInfo plane_info[AV_DRM_MAX_PLANES];
for (int i = 0; i < desc->nb_layers; i++) {
if (drm_to_vulkan_fmt(desc->layers[i].format) == VK_FORMAT_UNDEFINED) {
av_log(ctx, AV_LOG_ERROR, "Unsupported DMABUF layer format %#08x!\n",
desc->layers[i].format);
return AVERROR(EINVAL);
}
}
if (!(f = av_vk_frame_alloc())) {
av_log(ctx, AV_LOG_ERROR, "Unable to allocate memory for AVVkFrame!\n");
err = AVERROR(ENOMEM);
goto fail;
}
f->tiling = VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT;
for (int i = 0; i < desc->nb_layers; i++) {
const int planes = desc->layers[i].nb_planes;
/* Semaphore */
VkSemaphoreTypeCreateInfo sem_type_info = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_TYPE_CREATE_INFO,
.semaphoreType = VK_SEMAPHORE_TYPE_TIMELINE,
.initialValue = 0,
};
VkSemaphoreCreateInfo sem_spawn = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO,
.pNext = &sem_type_info,
};
/* Image creation */
VkSubresourceLayout ext_img_layouts[AV_DRM_MAX_PLANES];
VkImageDrmFormatModifierExplicitCreateInfoEXT ext_img_mod_spec = {
.sType = VK_STRUCTURE_TYPE_IMAGE_DRM_FORMAT_MODIFIER_EXPLICIT_CREATE_INFO_EXT,
.drmFormatModifier = desc->objects[0].format_modifier,
.drmFormatModifierPlaneCount = planes,
.pPlaneLayouts = (const VkSubresourceLayout *)&ext_img_layouts,
};
VkExternalMemoryImageCreateInfo ext_img_spec = {
.sType = VK_STRUCTURE_TYPE_EXTERNAL_MEMORY_IMAGE_CREATE_INFO,
.pNext = &ext_img_mod_spec,
.handleTypes = VK_EXTERNAL_MEMORY_HANDLE_TYPE_DMA_BUF_BIT_EXT,
};
VkImageCreateInfo create_info = {
.sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO,
.pNext = &ext_img_spec,
.imageType = VK_IMAGE_TYPE_2D,
.format = drm_to_vulkan_fmt(desc->layers[i].format),
.extent.depth = 1,
.mipLevels = 1,
.arrayLayers = 1,
.flags = 0x0,
.tiling = VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT,
.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED, /* specs say so */
.usage = VK_IMAGE_USAGE_SAMPLED_BIT |
VK_IMAGE_USAGE_TRANSFER_SRC_BIT,
.samples = VK_SAMPLE_COUNT_1_BIT,
.pQueueFamilyIndices = p->img_qfs,
.queueFamilyIndexCount = p->nb_img_qfs,
.sharingMode = p->nb_img_qfs > 1 ? VK_SHARING_MODE_CONCURRENT :
VK_SHARING_MODE_EXCLUSIVE,
};
/* Image format verification */
VkExternalImageFormatProperties ext_props = {
.sType = VK_STRUCTURE_TYPE_EXTERNAL_IMAGE_FORMAT_PROPERTIES_KHR,
};
VkImageFormatProperties2 props_ret = {
.sType = VK_STRUCTURE_TYPE_IMAGE_FORMAT_PROPERTIES_2,
.pNext = &ext_props,
};
VkPhysicalDeviceImageDrmFormatModifierInfoEXT props_drm_mod = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_IMAGE_DRM_FORMAT_MODIFIER_INFO_EXT,
.drmFormatModifier = ext_img_mod_spec.drmFormatModifier,
.pQueueFamilyIndices = create_info.pQueueFamilyIndices,
.queueFamilyIndexCount = create_info.queueFamilyIndexCount,
.sharingMode = create_info.sharingMode,
};
VkPhysicalDeviceExternalImageFormatInfo props_ext = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_EXTERNAL_IMAGE_FORMAT_INFO,
.pNext = &props_drm_mod,
.handleType = ext_img_spec.handleTypes,
};
VkPhysicalDeviceImageFormatInfo2 fmt_props = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_IMAGE_FORMAT_INFO_2,
.pNext = &props_ext,
.format = create_info.format,
.type = create_info.imageType,
.tiling = create_info.tiling,
.usage = create_info.usage,
.flags = create_info.flags,
};
/* Check if importing is possible for this combination of parameters */
ret = vk->GetPhysicalDeviceImageFormatProperties2(hwctx->phys_dev,
&fmt_props, &props_ret);
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Cannot map DRM frame to Vulkan: %s\n",
ff_vk_ret2str(ret));
err = AVERROR_EXTERNAL;
goto fail;
}
/* Set the image width/height */
get_plane_wh(&create_info.extent.width, &create_info.extent.height,
hwfc->sw_format, src->width, src->height, i);
/* Set the subresource layout based on the layer properties */
for (int j = 0; j < planes; j++) {
ext_img_layouts[j].offset = desc->layers[i].planes[j].offset;
ext_img_layouts[j].rowPitch = desc->layers[i].planes[j].pitch;
ext_img_layouts[j].size = 0; /* The specs say so for all 3 */
ext_img_layouts[j].arrayPitch = 0;
ext_img_layouts[j].depthPitch = 0;
}
/* Create image */
ret = vk->CreateImage(hwctx->act_dev, &create_info,
hwctx->alloc, &f->img[i]);
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Image creation failure: %s\n",
ff_vk_ret2str(ret));
err = AVERROR(EINVAL);
goto fail;
}
ret = vk->CreateSemaphore(hwctx->act_dev, &sem_spawn,
hwctx->alloc, &f->sem[i]);
if (ret != VK_SUCCESS) {
av_log(hwctx, AV_LOG_ERROR, "Failed to create semaphore: %s\n",
ff_vk_ret2str(ret));
return AVERROR_EXTERNAL;
}
/* We'd import a semaphore onto the one we created using
* vkImportSemaphoreFdKHR but unfortunately neither DRM nor VAAPI
* offer us anything we could import and sync with, so instead
* just signal the semaphore we created. */
f->queue_family[i] = p->nb_img_qfs > 1 ? VK_QUEUE_FAMILY_IGNORED : p->img_qfs[0];
f->layout[i] = create_info.initialLayout;
f->access[i] = 0x0;
f->sem_value[i] = 0;
}
for (int i = 0; i < desc->nb_layers; i++) {
/* Memory requirements */
VkImageMemoryRequirementsInfo2 req_desc = {
.sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_REQUIREMENTS_INFO_2,
.image = f->img[i],
};
VkMemoryDedicatedRequirements ded_req = {
.sType = VK_STRUCTURE_TYPE_MEMORY_DEDICATED_REQUIREMENTS,
};
VkMemoryRequirements2 req2 = {
.sType = VK_STRUCTURE_TYPE_MEMORY_REQUIREMENTS_2,
.pNext = &ded_req,
};
/* Allocation/importing */
VkMemoryFdPropertiesKHR fdmp = {
.sType = VK_STRUCTURE_TYPE_MEMORY_FD_PROPERTIES_KHR,
};
/* This assumes that a layer will never be constructed from multiple
* objects. If that was to happen in the real world, this code would
* need to import each plane separately.
*/
VkImportMemoryFdInfoKHR idesc = {
.sType = VK_STRUCTURE_TYPE_IMPORT_MEMORY_FD_INFO_KHR,
.fd = dup(desc->objects[desc->layers[i].planes[0].object_index].fd),
.handleType = VK_EXTERNAL_MEMORY_HANDLE_TYPE_DMA_BUF_BIT_EXT,
};
VkMemoryDedicatedAllocateInfo ded_alloc = {
.sType = VK_STRUCTURE_TYPE_MEMORY_DEDICATED_ALLOCATE_INFO,
.pNext = &idesc,
.image = req_desc.image,
};
/* Get object properties */
ret = vk->GetMemoryFdPropertiesKHR(hwctx->act_dev,
VK_EXTERNAL_MEMORY_HANDLE_TYPE_DMA_BUF_BIT_EXT,
idesc.fd, &fdmp);
if (ret != VK_SUCCESS) {
av_log(hwfc, AV_LOG_ERROR, "Failed to get FD properties: %s\n",
ff_vk_ret2str(ret));
err = AVERROR_EXTERNAL;
close(idesc.fd);
goto fail;
}
vk->GetImageMemoryRequirements2(hwctx->act_dev, &req_desc, &req2);
/* Only a single bit must be set, not a range, and it must match */
req2.memoryRequirements.memoryTypeBits = fdmp.memoryTypeBits;
err = alloc_mem(ctx, &req2.memoryRequirements,
VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT,
(ded_req.prefersDedicatedAllocation ||
ded_req.requiresDedicatedAllocation) ?
&ded_alloc : ded_alloc.pNext,
&f->flags, &f->mem[i]);
if (err) {
close(idesc.fd);
return err;
}
f->size[i] = req2.memoryRequirements.size;
}
for (int i = 0; i < desc->nb_layers; i++) {
const int planes = desc->layers[i].nb_planes;
for (int j = 0; j < planes; j++) {
VkImageAspectFlagBits aspect = j == 0 ? VK_IMAGE_ASPECT_MEMORY_PLANE_0_BIT_EXT :
j == 1 ? VK_IMAGE_ASPECT_MEMORY_PLANE_1_BIT_EXT :
VK_IMAGE_ASPECT_MEMORY_PLANE_2_BIT_EXT;
plane_info[bind_counts].sType = VK_STRUCTURE_TYPE_BIND_IMAGE_PLANE_MEMORY_INFO;
plane_info[bind_counts].pNext = NULL;
plane_info[bind_counts].planeAspect = aspect;
bind_info[bind_counts].sType = VK_STRUCTURE_TYPE_BIND_IMAGE_MEMORY_INFO;
bind_info[bind_counts].pNext = planes > 1 ? &plane_info[bind_counts] : NULL;
bind_info[bind_counts].image = f->img[i];
bind_info[bind_counts].memory = f->mem[i];
/* Offset is already signalled via pPlaneLayouts above */
bind_info[bind_counts].memoryOffset = 0;
bind_counts++;
}
}
/* Bind the allocated memory to the images */
ret = vk->BindImageMemory2(hwctx->act_dev, bind_counts, bind_info);
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Failed to bind memory: %s\n",
ff_vk_ret2str(ret));
err = AVERROR_EXTERNAL;
goto fail;
}
err = prepare_frame(hwfc, &fp->compute_exec, f, PREP_MODE_EXTERNAL_IMPORT);
if (err)
goto fail;
*frame = f;
return 0;
fail:
for (int i = 0; i < desc->nb_layers; i++) {
vk->DestroyImage(hwctx->act_dev, f->img[i], hwctx->alloc);
vk->DestroySemaphore(hwctx->act_dev, f->sem[i], hwctx->alloc);
}
for (int i = 0; i < desc->nb_objects; i++)
vk->FreeMemory(hwctx->act_dev, f->mem[i], hwctx->alloc);
av_free(f);
return err;
}
static int vulkan_map_from_drm(AVHWFramesContext *hwfc, AVFrame *dst,
const AVFrame *src, int flags)
{
int err = 0;
AVVkFrame *f;
if ((err = vulkan_map_from_drm_frame_desc(hwfc, &f, src)))
return err;
/* The unmapping function will free this */
dst->data[0] = (uint8_t *)f;
dst->width = src->width;
dst->height = src->height;
err = ff_hwframe_map_create(dst->hw_frames_ctx, dst, src,
&vulkan_unmap_from_drm, f);
if (err < 0)
goto fail;
av_log(hwfc, AV_LOG_DEBUG, "Mapped DRM object to Vulkan!\n");
return 0;
fail:
vulkan_frame_free(hwfc->device_ctx->hwctx, (uint8_t *)f);
dst->data[0] = NULL;
return err;
}
#if CONFIG_VAAPI
static int vulkan_map_from_vaapi(AVHWFramesContext *dst_fc,
AVFrame *dst, const AVFrame *src,
int flags)
{
int err;
AVFrame *tmp = av_frame_alloc();
AVHWFramesContext *vaapi_fc = (AVHWFramesContext*)src->hw_frames_ctx->data;
AVVAAPIDeviceContext *vaapi_ctx = vaapi_fc->device_ctx->hwctx;
VASurfaceID surface_id = (VASurfaceID)(uintptr_t)src->data[3];
if (!tmp)
return AVERROR(ENOMEM);
/* We have to sync since like the previous comment said, no semaphores */
vaSyncSurface(vaapi_ctx->display, surface_id);
tmp->format = AV_PIX_FMT_DRM_PRIME;
err = av_hwframe_map(tmp, src, flags);
if (err < 0)
goto fail;
err = vulkan_map_from_drm(dst_fc, dst, tmp, flags);
if (err < 0)
goto fail;
err = ff_hwframe_map_replace(dst, src);
fail:
av_frame_free(&tmp);
return err;
}
#endif
#endif
#if CONFIG_CUDA
static int vulkan_export_to_cuda(AVHWFramesContext *hwfc,
AVBufferRef *cuda_hwfc,
const AVFrame *frame)
{
int err;
VkResult ret;
AVVkFrame *dst_f;
AVVkFrameInternal *dst_int;
AVHWDeviceContext *ctx = hwfc->device_ctx;
AVVulkanDeviceContext *hwctx = ctx->hwctx;
const int planes = av_pix_fmt_count_planes(hwfc->sw_format);
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(hwfc->sw_format);
VulkanDevicePriv *p = ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVHWFramesContext *cuda_fc = (AVHWFramesContext*)cuda_hwfc->data;
AVHWDeviceContext *cuda_cu = cuda_fc->device_ctx;
AVCUDADeviceContext *cuda_dev = cuda_cu->hwctx;
AVCUDADeviceContextInternal *cu_internal = cuda_dev->internal;
CudaFunctions *cu = cu_internal->cuda_dl;
CUarray_format cufmt = desc->comp[0].depth > 8 ? CU_AD_FORMAT_UNSIGNED_INT16 :
CU_AD_FORMAT_UNSIGNED_INT8;
dst_f = (AVVkFrame *)frame->data[0];
dst_int = dst_f->internal;
if (!dst_int->cuda_fc_ref) {
dst_int->cuda_fc_ref = av_buffer_ref(cuda_hwfc);
if (!dst_int->cuda_fc_ref)
return AVERROR(ENOMEM);
for (int i = 0; i < planes; i++) {
CUDA_EXTERNAL_MEMORY_MIPMAPPED_ARRAY_DESC tex_desc = {
.offset = 0,
.arrayDesc = {
.Depth = 0,
.Format = cufmt,
.NumChannels = 1 + ((planes == 2) && i),
.Flags = 0,
},
.numLevels = 1,
};
int p_w, p_h;
#ifdef _WIN32
CUDA_EXTERNAL_MEMORY_HANDLE_DESC ext_desc = {
.type = IsWindows8OrGreater()
? CU_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32
: CU_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32_KMT,
.size = dst_f->size[i],
};
VkMemoryGetWin32HandleInfoKHR export_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_GET_WIN32_HANDLE_INFO_KHR,
.memory = dst_f->mem[i],
.handleType = IsWindows8OrGreater()
? VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32_BIT
: VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32_KMT_BIT,
};
VkSemaphoreGetWin32HandleInfoKHR sem_export = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_GET_WIN32_HANDLE_INFO_KHR,
.semaphore = dst_f->sem[i],
.handleType = IsWindows8OrGreater()
? VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_BIT
: VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_KMT_BIT,
};
CUDA_EXTERNAL_SEMAPHORE_HANDLE_DESC ext_sem_desc = {
.type = 10 /* TODO: CU_EXTERNAL_SEMAPHORE_HANDLE_TYPE_TIMELINE_SEMAPHORE_WIN32 */,
};
ret = vk->GetMemoryWin32HandleKHR(hwctx->act_dev, &export_info,
&ext_desc.handle.win32.handle);
if (ret != VK_SUCCESS) {
av_log(hwfc, AV_LOG_ERROR, "Unable to export the image as a Win32 Handle: %s!\n",
ff_vk_ret2str(ret));
err = AVERROR_EXTERNAL;
goto fail;
}
dst_int->ext_mem_handle[i] = ext_desc.handle.win32.handle;
#else
CUDA_EXTERNAL_MEMORY_HANDLE_DESC ext_desc = {
.type = CU_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD,
.size = dst_f->size[i],
};
VkMemoryGetFdInfoKHR export_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_GET_FD_INFO_KHR,
.memory = dst_f->mem[i],
.handleType = VK_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD_BIT_KHR,
};
VkSemaphoreGetFdInfoKHR sem_export = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_GET_FD_INFO_KHR,
.semaphore = dst_f->sem[i],
.handleType = VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_FD_BIT,
};
CUDA_EXTERNAL_SEMAPHORE_HANDLE_DESC ext_sem_desc = {
.type = 9 /* TODO: CU_EXTERNAL_SEMAPHORE_HANDLE_TYPE_TIMELINE_SEMAPHORE_FD */,
};
ret = vk->GetMemoryFdKHR(hwctx->act_dev, &export_info,
&ext_desc.handle.fd);
if (ret != VK_SUCCESS) {
av_log(hwfc, AV_LOG_ERROR, "Unable to export the image as a FD: %s!\n",
ff_vk_ret2str(ret));
err = AVERROR_EXTERNAL;
goto fail;
}
#endif
ret = CHECK_CU(cu->cuImportExternalMemory(&dst_int->ext_mem[i], &ext_desc));
if (ret < 0) {
#ifndef _WIN32
close(ext_desc.handle.fd);
#endif
err = AVERROR_EXTERNAL;
goto fail;
}
get_plane_wh(&p_w, &p_h, hwfc->sw_format, hwfc->width, hwfc->height, i);
tex_desc.arrayDesc.Width = p_w;
tex_desc.arrayDesc.Height = p_h;
ret = CHECK_CU(cu->cuExternalMemoryGetMappedMipmappedArray(&dst_int->cu_mma[i],
dst_int->ext_mem[i],
&tex_desc));
if (ret < 0) {
err = AVERROR_EXTERNAL;
goto fail;
}
ret = CHECK_CU(cu->cuMipmappedArrayGetLevel(&dst_int->cu_array[i],
dst_int->cu_mma[i], 0));
if (ret < 0) {
err = AVERROR_EXTERNAL;
goto fail;
}
#ifdef _WIN32
ret = vk->GetSemaphoreWin32HandleKHR(hwctx->act_dev, &sem_export,
&ext_sem_desc.handle.win32.handle);
#else
ret = vk->GetSemaphoreFdKHR(hwctx->act_dev, &sem_export,
&ext_sem_desc.handle.fd);
#endif
if (ret != VK_SUCCESS) {
av_log(ctx, AV_LOG_ERROR, "Failed to export semaphore: %s\n",
ff_vk_ret2str(ret));
err = AVERROR_EXTERNAL;
goto fail;
}
#ifdef _WIN32
dst_int->ext_sem_handle[i] = ext_sem_desc.handle.win32.handle;
#endif
ret = CHECK_CU(cu->cuImportExternalSemaphore(&dst_int->cu_sem[i],
&ext_sem_desc));
if (ret < 0) {
#ifndef _WIN32
close(ext_sem_desc.handle.fd);
#endif
err = AVERROR_EXTERNAL;
goto fail;
}
}
}
return 0;
fail:
vulkan_free_internal(dst_f);
return err;
}
static int vulkan_transfer_data_from_cuda(AVHWFramesContext *hwfc,
AVFrame *dst, const AVFrame *src)
{
int err;
CUcontext dummy;
AVVkFrame *dst_f;
AVVkFrameInternal *dst_int;
VulkanFramesPriv *fp = hwfc->internal->priv;
const int planes = av_pix_fmt_count_planes(hwfc->sw_format);
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(hwfc->sw_format);
AVHWFramesContext *cuda_fc = (AVHWFramesContext*)src->hw_frames_ctx->data;
AVHWDeviceContext *cuda_cu = cuda_fc->device_ctx;
AVCUDADeviceContext *cuda_dev = cuda_cu->hwctx;
AVCUDADeviceContextInternal *cu_internal = cuda_dev->internal;
CudaFunctions *cu = cu_internal->cuda_dl;
CUDA_EXTERNAL_SEMAPHORE_WAIT_PARAMS s_w_par[AV_NUM_DATA_POINTERS] = { 0 };
CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS s_s_par[AV_NUM_DATA_POINTERS] = { 0 };
dst_f = (AVVkFrame *)dst->data[0];
err = prepare_frame(hwfc, &fp->upload_exec, dst_f, PREP_MODE_EXTERNAL_EXPORT);
if (err < 0)
return err;
err = CHECK_CU(cu->cuCtxPushCurrent(cuda_dev->cuda_ctx));
if (err < 0)
return err;
err = vulkan_export_to_cuda(hwfc, src->hw_frames_ctx, dst);
if (err < 0) {
CHECK_CU(cu->cuCtxPopCurrent(&dummy));
return err;
}
dst_int = dst_f->internal;
for (int i = 0; i < planes; i++) {
s_w_par[i].params.fence.value = dst_f->sem_value[i] + 0;
s_s_par[i].params.fence.value = dst_f->sem_value[i] + 1;
}
err = CHECK_CU(cu->cuWaitExternalSemaphoresAsync(dst_int->cu_sem, s_w_par,
planes, cuda_dev->stream));
if (err < 0)
goto fail;
for (int i = 0; i < planes; i++) {
CUDA_MEMCPY2D cpy = {
.srcMemoryType = CU_MEMORYTYPE_DEVICE,
.srcDevice = (CUdeviceptr)src->data[i],
.srcPitch = src->linesize[i],
.srcY = 0,
.dstMemoryType = CU_MEMORYTYPE_ARRAY,
.dstArray = dst_int->cu_array[i],
};
int p_w, p_h;
get_plane_wh(&p_w, &p_h, hwfc->sw_format, hwfc->width, hwfc->height, i);
cpy.WidthInBytes = p_w * desc->comp[i].step;
cpy.Height = p_h;
err = CHECK_CU(cu->cuMemcpy2DAsync(&cpy, cuda_dev->stream));
if (err < 0)
goto fail;
}
err = CHECK_CU(cu->cuSignalExternalSemaphoresAsync(dst_int->cu_sem, s_s_par,
planes, cuda_dev->stream));
if (err < 0)
goto fail;
for (int i = 0; i < planes; i++)
dst_f->sem_value[i]++;
CHECK_CU(cu->cuCtxPopCurrent(&dummy));
av_log(hwfc, AV_LOG_VERBOSE, "Transfered CUDA image to Vulkan!\n");
return err = prepare_frame(hwfc, &fp->upload_exec, dst_f, PREP_MODE_EXTERNAL_IMPORT);
fail:
CHECK_CU(cu->cuCtxPopCurrent(&dummy));
vulkan_free_internal(dst_f);
dst_f->internal = NULL;
av_buffer_unref(&dst->buf[0]);
return err;
}
#endif
static int vulkan_map_to(AVHWFramesContext *hwfc, AVFrame *dst,
const AVFrame *src, int flags)
{
av_unused VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
switch (src->format) {
#if CONFIG_LIBDRM
#if CONFIG_VAAPI
case AV_PIX_FMT_VAAPI:
if (p->vkctx.extensions & FF_VK_EXT_DRM_MODIFIER_FLAGS)
return vulkan_map_from_vaapi(hwfc, dst, src, flags);
else
return AVERROR(ENOSYS);
#endif
case AV_PIX_FMT_DRM_PRIME:
if (p->vkctx.extensions & FF_VK_EXT_DRM_MODIFIER_FLAGS)
return vulkan_map_from_drm(hwfc, dst, src, flags);
else
return AVERROR(ENOSYS);
#endif
default:
return AVERROR(ENOSYS);
}
}
#if CONFIG_LIBDRM
typedef struct VulkanDRMMapping {
AVDRMFrameDescriptor drm_desc;
AVVkFrame *source;
} VulkanDRMMapping;
static void vulkan_unmap_to_drm(AVHWFramesContext *hwfc, HWMapDescriptor *hwmap)
{
AVDRMFrameDescriptor *drm_desc = hwmap->priv;
for (int i = 0; i < drm_desc->nb_objects; i++)
close(drm_desc->objects[i].fd);
av_free(drm_desc);
}
static inline uint32_t vulkan_fmt_to_drm(VkFormat vkfmt)
{
for (int i = 0; i < FF_ARRAY_ELEMS(vulkan_drm_format_map); i++)
if (vulkan_drm_format_map[i].vk_format == vkfmt)
return vulkan_drm_format_map[i].drm_fourcc;
return DRM_FORMAT_INVALID;
}
static int vulkan_map_to_drm(AVHWFramesContext *hwfc, AVFrame *dst,
const AVFrame *src, int flags)
{
int err = 0;
VkResult ret;
AVVkFrame *f = (AVVkFrame *)src->data[0];
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
VulkanFramesPriv *fp = hwfc->internal->priv;
AVVulkanDeviceContext *hwctx = hwfc->device_ctx->hwctx;
AVVulkanFramesContext *hwfctx = hwfc->hwctx;
const int planes = av_pix_fmt_count_planes(hwfc->sw_format);
VkImageDrmFormatModifierPropertiesEXT drm_mod = {
.sType = VK_STRUCTURE_TYPE_IMAGE_DRM_FORMAT_MODIFIER_PROPERTIES_EXT,
};
VkSemaphoreWaitInfo wait_info = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_WAIT_INFO,
.flags = 0x0,
.semaphoreCount = planes,
};
AVDRMFrameDescriptor *drm_desc = av_mallocz(sizeof(*drm_desc));
if (!drm_desc)
return AVERROR(ENOMEM);
err = prepare_frame(hwfc, &fp->compute_exec, f, PREP_MODE_EXTERNAL_EXPORT);
if (err < 0)
goto end;
/* Wait for the operation to finish so we can cleanly export it. */
wait_info.pSemaphores = f->sem;
wait_info.pValues = f->sem_value;
vk->WaitSemaphores(hwctx->act_dev, &wait_info, UINT64_MAX);
err = ff_hwframe_map_create(src->hw_frames_ctx, dst, src, &vulkan_unmap_to_drm, drm_desc);
if (err < 0)
goto end;
ret = vk->GetImageDrmFormatModifierPropertiesEXT(hwctx->act_dev, f->img[0],
&drm_mod);
if (ret != VK_SUCCESS) {
av_log(hwfc, AV_LOG_ERROR, "Failed to retrieve DRM format modifier!\n");
err = AVERROR_EXTERNAL;
goto end;
}
for (int i = 0; (i < planes) && (f->mem[i]); i++) {
VkMemoryGetFdInfoKHR export_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_GET_FD_INFO_KHR,
.memory = f->mem[i],
.handleType = VK_EXTERNAL_MEMORY_HANDLE_TYPE_DMA_BUF_BIT_EXT,
};
ret = vk->GetMemoryFdKHR(hwctx->act_dev, &export_info,
&drm_desc->objects[i].fd);
if (ret != VK_SUCCESS) {
av_log(hwfc, AV_LOG_ERROR, "Unable to export the image as a FD!\n");
err = AVERROR_EXTERNAL;
goto end;
}
drm_desc->nb_objects++;
drm_desc->objects[i].size = f->size[i];
drm_desc->objects[i].format_modifier = drm_mod.drmFormatModifier;
}
drm_desc->nb_layers = planes;
for (int i = 0; i < drm_desc->nb_layers; i++) {
VkSubresourceLayout layout;
VkImageSubresource sub = {
.aspectMask = VK_IMAGE_ASPECT_MEMORY_PLANE_0_BIT_EXT,
};
VkFormat plane_vkfmt = av_vkfmt_from_pixfmt(hwfc->sw_format)[i];
drm_desc->layers[i].format = vulkan_fmt_to_drm(plane_vkfmt);
drm_desc->layers[i].nb_planes = 1;
if (drm_desc->layers[i].format == DRM_FORMAT_INVALID) {
av_log(hwfc, AV_LOG_ERROR, "Cannot map to DRM layer, unsupported!\n");
err = AVERROR_PATCHWELCOME;
goto end;
}
drm_desc->layers[i].planes[0].object_index = FFMIN(i, drm_desc->nb_objects - 1);
if (f->tiling == VK_IMAGE_TILING_OPTIMAL)
continue;
vk->GetImageSubresourceLayout(hwctx->act_dev, f->img[i], &sub, &layout);
drm_desc->layers[i].planes[0].offset = layout.offset;
drm_desc->layers[i].planes[0].pitch = layout.rowPitch;
if (hwfctx->flags & AV_VK_FRAME_FLAG_CONTIGUOUS_MEMORY)
drm_desc->layers[i].planes[0].offset += f->offset[i];
}
dst->width = src->width;
dst->height = src->height;
dst->data[0] = (uint8_t *)drm_desc;
av_log(hwfc, AV_LOG_VERBOSE, "Mapped AVVkFrame to a DRM object!\n");
return 0;
end:
av_free(drm_desc);
return err;
}
#if CONFIG_VAAPI
static int vulkan_map_to_vaapi(AVHWFramesContext *hwfc, AVFrame *dst,
const AVFrame *src, int flags)
{
int err;
AVFrame *tmp = av_frame_alloc();
if (!tmp)
return AVERROR(ENOMEM);
tmp->format = AV_PIX_FMT_DRM_PRIME;
err = vulkan_map_to_drm(hwfc, tmp, src, flags);
if (err < 0)
goto fail;
err = av_hwframe_map(dst, tmp, flags);
if (err < 0)
goto fail;
err = ff_hwframe_map_replace(dst, src);
fail:
av_frame_free(&tmp);
return err;
}
#endif
#endif
static int vulkan_map_from(AVHWFramesContext *hwfc, AVFrame *dst,
const AVFrame *src, int flags)
{
av_unused VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
switch (dst->format) {
#if CONFIG_LIBDRM
case AV_PIX_FMT_DRM_PRIME:
if (p->vkctx.extensions & FF_VK_EXT_DRM_MODIFIER_FLAGS)
return vulkan_map_to_drm(hwfc, dst, src, flags);
else
return AVERROR(ENOSYS);
#if CONFIG_VAAPI
case AV_PIX_FMT_VAAPI:
if (p->vkctx.extensions & FF_VK_EXT_DRM_MODIFIER_FLAGS)
return vulkan_map_to_vaapi(hwfc, dst, src, flags);
else
return AVERROR(ENOSYS);
#endif
#endif
default:
break;
}
return AVERROR(ENOSYS);
}
static size_t get_req_buffer_size(VulkanDevicePriv *p, int *stride, int height)
{
size_t size;
*stride = FFALIGN(*stride, p->props.properties.limits.optimalBufferCopyRowPitchAlignment);
size = height*(*stride);
size = FFALIGN(size, p->props.properties.limits.minMemoryMapAlignment);
return size;
}
static int transfer_image_buf(AVHWFramesContext *hwfc, AVFrame *f,
AVBufferRef **bufs, size_t *buf_offsets,
const int *buf_stride, int w,
int h, enum AVPixelFormat pix_fmt, int to_buf)
{
int err;
AVVkFrame *frame = (AVVkFrame *)f->data[0];
VulkanFramesPriv *fp = hwfc->internal->priv;
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
VkImageMemoryBarrier2 img_bar[AV_NUM_DATA_POINTERS];
int nb_img_bar = 0;
const int nb_images = ff_vk_count_images(frame);
int pixfmt_planes = av_pix_fmt_count_planes(pix_fmt);
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt);
VkCommandBuffer cmd_buf;
FFVkExecContext *exec = ff_vk_exec_get(to_buf ? &fp->download_exec :
&fp->upload_exec);
cmd_buf = exec->buf;
ff_vk_exec_start(&p->vkctx, exec);
err = ff_vk_exec_add_dep_buf(&p->vkctx, exec, bufs, pixfmt_planes, 1);
if (err < 0)
return err;
err = ff_vk_exec_add_dep_frame(&p->vkctx, exec, f,
VK_PIPELINE_STAGE_2_ALL_COMMANDS_BIT,
VK_PIPELINE_STAGE_2_TRANSFER_BIT);
if (err < 0)
return err;
ff_vk_frame_barrier(&p->vkctx, exec, f, img_bar, &nb_img_bar,
VK_PIPELINE_STAGE_2_ALL_COMMANDS_BIT,
VK_PIPELINE_STAGE_2_TRANSFER_BIT_KHR,
to_buf ? VK_ACCESS_TRANSFER_READ_BIT :
VK_ACCESS_TRANSFER_WRITE_BIT,
to_buf ? VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL :
VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL,
VK_QUEUE_FAMILY_IGNORED);
vk->CmdPipelineBarrier2(cmd_buf, &(VkDependencyInfo) {
.sType = VK_STRUCTURE_TYPE_DEPENDENCY_INFO,
.pImageMemoryBarriers = img_bar,
.imageMemoryBarrierCount = nb_img_bar,
});
/* Schedule a copy for each plane */
for (int i = 0; i < pixfmt_planes; i++) {
int idx = FFMIN(i, nb_images - 1);
VkImageAspectFlags plane_aspect[] = { VK_IMAGE_ASPECT_COLOR_BIT,
VK_IMAGE_ASPECT_PLANE_0_BIT,
VK_IMAGE_ASPECT_PLANE_1_BIT,
VK_IMAGE_ASPECT_PLANE_2_BIT, };
FFVkBuffer *vkbuf = (FFVkBuffer *)bufs[i]->data;
VkBufferImageCopy buf_reg = {
.bufferOffset = buf_offsets[i],
.bufferRowLength = buf_stride[i] / desc->comp[i].step,
.imageSubresource.layerCount = 1,
.imageSubresource.aspectMask = plane_aspect[(pixfmt_planes != nb_images) +
i*(pixfmt_planes != nb_images)],
.imageOffset = { 0, 0, 0, },
};
uint32_t p_w, p_h;
get_plane_wh(&p_w, &p_h, pix_fmt, w, h, i);
buf_reg.bufferImageHeight = p_h;
buf_reg.imageExtent = (VkExtent3D){ p_w, p_h, 1, };
if (to_buf)
vk->CmdCopyImageToBuffer(cmd_buf, frame->img[idx],
img_bar[0].newLayout,
vkbuf->buf,
1, &buf_reg);
else
vk->CmdCopyBufferToImage(cmd_buf, vkbuf->buf, frame->img[idx],
img_bar[0].newLayout,
1, &buf_reg);
}
err = ff_vk_exec_submit(&p->vkctx, exec);
if (err < 0)
return err;
/* Wait for the operation to complete when downloading */
if (to_buf)
ff_vk_exec_wait(&p->vkctx, exec);
return 0;
}
static int vulkan_transfer_data(AVHWFramesContext *hwfc, const AVFrame *vkf,
const AVFrame *swf, int from)
{
int err = 0;
VkResult ret;
AVHWDeviceContext *dev_ctx = hwfc->device_ctx;
AVVulkanDeviceContext *hwctx = dev_ctx->hwctx;
VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
FFVulkanFunctions *vk = &p->vkctx.vkfn;
AVFrame tmp;
FFVkBuffer *vkbufs[AV_NUM_DATA_POINTERS];
AVBufferRef *bufs[AV_NUM_DATA_POINTERS] = { 0 };
size_t buf_offsets[AV_NUM_DATA_POINTERS] = { 0 };
uint32_t p_w, p_h;
const int planes = av_pix_fmt_count_planes(swf->format);
int host_mapped[AV_NUM_DATA_POINTERS] = { 0 };
const int map_host = !!(p->vkctx.extensions & FF_VK_EXT_EXTERNAL_HOST_MEMORY);
if ((swf->format != AV_PIX_FMT_NONE && !av_vkfmt_from_pixfmt(swf->format))) {
av_log(hwfc, AV_LOG_ERROR, "Unsupported software frame pixel format!\n");
return AVERROR(EINVAL);
}
if (swf->width > hwfc->width || swf->height > hwfc->height)
return AVERROR(EINVAL);
/* Create buffers */
for (int i = 0; i < planes; i++) {
size_t req_size;
VkExternalMemoryBufferCreateInfo create_desc = {
.sType = VK_STRUCTURE_TYPE_EXTERNAL_MEMORY_BUFFER_CREATE_INFO,
.handleTypes = VK_EXTERNAL_MEMORY_HANDLE_TYPE_HOST_ALLOCATION_BIT_EXT,
};
VkImportMemoryHostPointerInfoEXT import_desc = {
.sType = VK_STRUCTURE_TYPE_IMPORT_MEMORY_HOST_POINTER_INFO_EXT,
.handleType = VK_EXTERNAL_MEMORY_HANDLE_TYPE_HOST_ALLOCATION_BIT_EXT,
};
VkMemoryHostPointerPropertiesEXT p_props = {
.sType = VK_STRUCTURE_TYPE_MEMORY_HOST_POINTER_PROPERTIES_EXT,
};
get_plane_wh(&p_w, &p_h, swf->format, swf->width, swf->height, i);
tmp.linesize[i] = FFABS(swf->linesize[i]);
/* Do not map images with a negative stride */
if (map_host && swf->linesize[i] > 0) {
size_t offs;
offs = (uintptr_t)swf->data[i] % p->hprops.minImportedHostPointerAlignment;
import_desc.pHostPointer = swf->data[i] - offs;
/* We have to compensate for the few extra bytes of padding we
* completely ignore at the start */
req_size = FFALIGN(offs + tmp.linesize[i] * p_h,
p->hprops.minImportedHostPointerAlignment);
ret = vk->GetMemoryHostPointerPropertiesEXT(hwctx->act_dev,
import_desc.handleType,
import_desc.pHostPointer,
&p_props);
if (ret == VK_SUCCESS && p_props.memoryTypeBits) {
host_mapped[i] = 1;
buf_offsets[i] = offs;
}
}
if (!host_mapped[i])
req_size = get_req_buffer_size(p, &tmp.linesize[i], p_h);
err = ff_vk_create_avbuf(&p->vkctx, &bufs[i], req_size,
host_mapped[i] ? &create_desc : NULL,
host_mapped[i] ? &import_desc : NULL,
from ? VK_BUFFER_USAGE_TRANSFER_DST_BIT :
VK_BUFFER_USAGE_TRANSFER_SRC_BIT,
VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT |
(host_mapped[i] ?
VK_MEMORY_PROPERTY_HOST_COHERENT_BIT : 0x0));
if (err < 0)
goto end;
vkbufs[i] = (FFVkBuffer *)bufs[i]->data;
}
if (!from) {
/* Map, copy image TO buffer (which then goes to the VkImage), unmap */
if ((err = ff_vk_map_buffers(&p->vkctx, vkbufs, tmp.data, planes, 0)))
goto end;
for (int i = 0; i < planes; i++) {
if (host_mapped[i])
continue;
get_plane_wh(&p_w, &p_h, swf->format, swf->width, swf->height, i);
av_image_copy_plane(tmp.data[i], tmp.linesize[i],
(const uint8_t *)swf->data[i], swf->linesize[i],
FFMIN(tmp.linesize[i], FFABS(swf->linesize[i])),
p_h);
}
if ((err = ff_vk_unmap_buffers(&p->vkctx, vkbufs, planes, 1)))
goto end;
}
/* Copy buffers into/from image */
err = transfer_image_buf(hwfc, (AVFrame *)vkf, bufs, buf_offsets,
tmp.linesize, swf->width, swf->height, swf->format,
from);
if (from) {
/* Map, copy buffer (which came FROM the VkImage) to the frame, unmap */
if ((err = ff_vk_map_buffers(&p->vkctx, vkbufs, tmp.data, planes, 0)))
goto end;
for (int i = 0; i < planes; i++) {
if (host_mapped[i])
continue;
get_plane_wh(&p_w, &p_h, swf->format, swf->width, swf->height, i);
av_image_copy_plane_uc_from(swf->data[i], swf->linesize[i],
(const uint8_t *)tmp.data[i], tmp.linesize[i],
FFMIN(tmp.linesize[i], FFABS(swf->linesize[i])),
p_h);
}
if ((err = ff_vk_unmap_buffers(&p->vkctx, vkbufs, planes, 1)))
goto end;
}
end:
for (int i = 0; i < planes; i++)
av_buffer_unref(&bufs[i]);
return err;
}
static int vulkan_transfer_data_to(AVHWFramesContext *hwfc, AVFrame *dst,
const AVFrame *src)
{
av_unused VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
switch (src->format) {
#if CONFIG_CUDA
case AV_PIX_FMT_CUDA:
#ifdef _WIN32
if ((p->vkctx.extensions & FF_VK_EXT_EXTERNAL_WIN32_MEMORY) &&
(p->vkctx.extensions & FF_VK_EXT_EXTERNAL_WIN32_SEM))
#else
if ((p->vkctx.extensions & FF_VK_EXT_EXTERNAL_FD_MEMORY) &&
(p->vkctx.extensions & FF_VK_EXT_EXTERNAL_FD_SEM))
#endif
return vulkan_transfer_data_from_cuda(hwfc, dst, src);
#endif
default:
if (src->hw_frames_ctx)
return AVERROR(ENOSYS);
else
return vulkan_transfer_data(hwfc, dst, src, 0);
}
}
#if CONFIG_CUDA
static int vulkan_transfer_data_to_cuda(AVHWFramesContext *hwfc, AVFrame *dst,
const AVFrame *src)
{
int err;
CUcontext dummy;
AVVkFrame *dst_f;
AVVkFrameInternal *dst_int;
VulkanFramesPriv *fp = hwfc->internal->priv;
const int planes = av_pix_fmt_count_planes(hwfc->sw_format);
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(hwfc->sw_format);
AVHWFramesContext *cuda_fc = (AVHWFramesContext*)dst->hw_frames_ctx->data;
AVHWDeviceContext *cuda_cu = cuda_fc->device_ctx;
AVCUDADeviceContext *cuda_dev = cuda_cu->hwctx;
AVCUDADeviceContextInternal *cu_internal = cuda_dev->internal;
CudaFunctions *cu = cu_internal->cuda_dl;
CUDA_EXTERNAL_SEMAPHORE_WAIT_PARAMS s_w_par[AV_NUM_DATA_POINTERS] = { 0 };
CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS s_s_par[AV_NUM_DATA_POINTERS] = { 0 };
dst_f = (AVVkFrame *)src->data[0];
err = prepare_frame(hwfc, &fp->upload_exec, dst_f, PREP_MODE_EXTERNAL_EXPORT);
if (err < 0)
return err;
err = CHECK_CU(cu->cuCtxPushCurrent(cuda_dev->cuda_ctx));
if (err < 0)
return err;
err = vulkan_export_to_cuda(hwfc, dst->hw_frames_ctx, src);
if (err < 0) {
CHECK_CU(cu->cuCtxPopCurrent(&dummy));
return err;
}
dst_int = dst_f->internal;
for (int i = 0; i < planes; i++) {
s_w_par[i].params.fence.value = dst_f->sem_value[i] + 0;
s_s_par[i].params.fence.value = dst_f->sem_value[i] + 1;
}
err = CHECK_CU(cu->cuWaitExternalSemaphoresAsync(dst_int->cu_sem, s_w_par,
planes, cuda_dev->stream));
if (err < 0)
goto fail;
for (int i = 0; i < planes; i++) {
CUDA_MEMCPY2D cpy = {
.dstMemoryType = CU_MEMORYTYPE_DEVICE,
.dstDevice = (CUdeviceptr)dst->data[i],
.dstPitch = dst->linesize[i],
.dstY = 0,
.srcMemoryType = CU_MEMORYTYPE_ARRAY,
.srcArray = dst_int->cu_array[i],
};
int w, h;
get_plane_wh(&w, &h, hwfc->sw_format, hwfc->width, hwfc->height, i);
cpy.WidthInBytes = w * desc->comp[i].step;
cpy.Height = h;
err = CHECK_CU(cu->cuMemcpy2DAsync(&cpy, cuda_dev->stream));
if (err < 0)
goto fail;
}
err = CHECK_CU(cu->cuSignalExternalSemaphoresAsync(dst_int->cu_sem, s_s_par,
planes, cuda_dev->stream));
if (err < 0)
goto fail;
for (int i = 0; i < planes; i++)
dst_f->sem_value[i]++;
CHECK_CU(cu->cuCtxPopCurrent(&dummy));
av_log(hwfc, AV_LOG_VERBOSE, "Transfered Vulkan image to CUDA!\n");
return prepare_frame(hwfc, &fp->upload_exec, dst_f, PREP_MODE_EXTERNAL_IMPORT);
fail:
CHECK_CU(cu->cuCtxPopCurrent(&dummy));
vulkan_free_internal(dst_f);
dst_f->internal = NULL;
av_buffer_unref(&dst->buf[0]);
return err;
}
#endif
static int vulkan_transfer_data_from(AVHWFramesContext *hwfc, AVFrame *dst,
const AVFrame *src)
{
av_unused VulkanDevicePriv *p = hwfc->device_ctx->internal->priv;
switch (dst->format) {
#if CONFIG_CUDA
case AV_PIX_FMT_CUDA:
#ifdef _WIN32
if ((p->vkctx.extensions & FF_VK_EXT_EXTERNAL_WIN32_MEMORY) &&
(p->vkctx.extensions & FF_VK_EXT_EXTERNAL_WIN32_SEM))
#else
if ((p->vkctx.extensions & FF_VK_EXT_EXTERNAL_FD_MEMORY) &&
(p->vkctx.extensions & FF_VK_EXT_EXTERNAL_FD_SEM))
#endif
return vulkan_transfer_data_to_cuda(hwfc, dst, src);
#endif
default:
if (dst->hw_frames_ctx)
return AVERROR(ENOSYS);
else
return vulkan_transfer_data(hwfc, src, dst, 1);
}
}
static int vulkan_frames_derive_to(AVHWFramesContext *dst_fc,
AVHWFramesContext *src_fc, int flags)
{
return vulkan_frames_init(dst_fc);
}
AVVkFrame *av_vk_frame_alloc(void)
{
int err;
AVVkFrame *f = av_mallocz(sizeof(AVVkFrame));
if (!f)
return NULL;
f->internal = av_mallocz(sizeof(*f->internal));
if (!f->internal) {
av_free(f);
return NULL;
}
err = pthread_mutex_init(&f->internal->update_mutex, NULL);
if (err != 0) {
av_free(f->internal);
av_free(f);
return NULL;
}
return f;
}
const HWContextType ff_hwcontext_type_vulkan = {
.type = AV_HWDEVICE_TYPE_VULKAN,
.name = "Vulkan",
.device_hwctx_size = sizeof(AVVulkanDeviceContext),
.device_priv_size = sizeof(VulkanDevicePriv),
.frames_hwctx_size = sizeof(AVVulkanFramesContext),
.frames_priv_size = sizeof(VulkanFramesPriv),
.device_init = &vulkan_device_init,
.device_create = &vulkan_device_create,
.device_derive = &vulkan_device_derive,
.frames_get_constraints = &vulkan_frames_get_constraints,
.frames_init = vulkan_frames_init,
.frames_get_buffer = vulkan_get_buffer,
.frames_uninit = vulkan_frames_uninit,
.transfer_get_formats = vulkan_transfer_get_formats,
.transfer_data_to = vulkan_transfer_data_to,
.transfer_data_from = vulkan_transfer_data_from,
.map_to = vulkan_map_to,
.map_from = vulkan_map_from,
.frames_derive_to = &vulkan_frames_derive_to,
.pix_fmts = (const enum AVPixelFormat []) {
AV_PIX_FMT_VULKAN,
AV_PIX_FMT_NONE
},
};
Reference in New Issue View Git Blame Copy Permalink