FFmpeg

Commit Graph

Author	SHA1	Message	Date
Niklas Haas	9675e54b02	avutil/hwcontext_vulkan: add libplacebo required features For compatibility with vf_libplacebo	1 year ago
Niklas Haas	b41ea8c6c3	avfilter/vf_libplacebo: bump max vk version For two reasons: 1. We now create a vulkan 1.3 device 2. libplacebo master currently requires a vulkan 1.3 device	1 year ago
Niklas Haas	661db605db	avfilter/vf_libplacebo: forward queue locking primitives For thread safety.	1 year ago
Lynne	f70f873272	transpose_vulkan: port for the rewrite	1 year ago
Lynne	d4b51b5085	scale_vulkan: port for the rewrite	1 year ago
Lynne	8e9ceb1efb	overlay_vulkan: port for the rewrite	1 year ago
Lynne	59f21c051d	gblur_vulkan: port for the rewrite	1 year ago
Lynne	d26cbae3e1	flip_vulkan: port for the rewrite	1 year ago
Lynne	125636608d	chromaber_vulkan: port for the rewrite	1 year ago
Lynne	40bf1d525e	blend_vulkan: port for the rewrite	1 year ago
Lynne	f29850e98f	avgblur_vulkan: port for the rewrite	1 year ago
Lynne	05ce6473ac	lavfi: add lavfi-only Vulkan infrastructure	1 year ago
Lynne	51b7fe81be	hwcontext_vulkan: enable additional device properties	1 year ago
Lynne	33fc919bb7	hwcontext_vulkan: remove duplicate code, port to use generic vulkan utils The temporary AVFrame on staack enables us to use the common dependency/dispatch code in prepare_frame(). The prepare_frame() function is used for both frame initialization and frame import/export queue family transfer operations. In the former case, no AVFrame exists yet, so, as this is purely libavutil code, we create a temporary frame on stack. Otherwise, we'd need to allocate multiple frames somewhere, one for each possible command buffer dispatch.	1 year ago
Lynne	94e17a63a4	hwcontext_vulkan: don't change properties if prepare_frame fails	1 year ago
Lynne	32fc36ee61	hwcontext_vulkan: remove linear+host_visible "fast" path The idea was that it's faster to map linear images and copy them via regular memcpy. This is a very niche use, plus very inconsistently useful, as it would only really be faster on a few Intel GPUs. Even then, using the non-cached memcpy would've been better. Instead, scrap this code. Drivers are better at figuring out what copy to use, and if we're host-mapping, it should actually be just as fast, if not faster.	1 year ago
Lynne	48f85de0e7	hwcontext_vulkan: rewrite to support multiplane surfaces This commit adds proper handling of multiplane images throughout all of the hwcontext code. To avoid breakage of individual components, the change is performed as a single commit.	1 year ago
Lynne	a4d63b46d9	vulkan: make GLSL macro functions semicolumn-safe	1 year ago
Lynne	83024beec2	vulkan: enable forcing of full subgroups	1 year ago
Lynne	758f8b26b9	vulkan: add ff_vk_count_images()	1 year ago
Lynne	b5eaeb1f13	vulkan: rewrite to support all necessary features This commit rewrites the majority of vulkan.c to enable its use as a general-purpose high-level utility code, usable for decoding, encoding, and filtering of video frames. The dependency system was rewritten to simplify management of execution. The image handling system was rewritten to accomodate multiplane images. Due to how related all the new features were, this is a single commit.	1 year ago
Lynne	721b71da4a	vulkan: return current queue index from ff_vk_qf_rotate()	1 year ago
Lynne	b15104ed97	vulkan: add support for retrieving queue, query and video properties	1 year ago
Lynne	6eaf3fe69c	vulkan: add support for queries	1 year ago
Lynne	f3fb1b50bb	vulkan: minor indent fix, add support for synchronous submission/waiting	1 year ago
Lynne	d386988c39	vulkan: use device properties 2 and add a convenience loader function	1 year ago
Lynne	bf69a64135	vulkan: add size tracking to buffer structs	1 year ago
Lynne	b18e20a4ee	vulkan: do not wait for device idle when destroying buffers This should be done explicitly.	1 year ago
Lynne	15de0af8f0	vulkan: allow alloc pNext in ff_vk_create_buf	1 year ago
Lynne	af48790465	vulkan: support ignoring memory properties when allocating	1 year ago
Lynne	3c2f43d8ee	vulkan: expose ff_vk_alloc_mem()	1 year ago
Lynne	fa67ccee37	vulkan: add ff_vk_image_create()	1 year ago
Lynne	e8fce74abf	vulkan: add ff_vk_qf_fill()	1 year ago
Lynne	b5e333bba7	vulkan: add pNext argument to ff_vk_create_buf()	1 year ago
Lynne	a0d47a2ad9	vulkan: fix comment statement about exec_queue blocking	1 year ago
Lynne	619b1265a2	vulkan: add additional error codes	1 year ago
Lynne	0c9c0e40fb	vulkan: define VK_NO_PROTOTYPES This just disables the vulkan headers from defining any symbols like vkCmdPipelineBarrier2(). Instead, all functions must be loaded via the loader and used as function pointers as vk->CmdPipelineBarrier2. Mostly just forces developers to write correct code, as using the symbols can be undesirable in case API users define their own function wrappers via the loader API.	1 year ago
Lynne	92ddd415bc	vulkan: lock queues before submitting operations	1 year ago
Lynne	9b385b480f	hwcontext_vulkan: enable GPU-assisted validation when debugging	1 year ago
Lynne	e5e12c5078	hwcontext_vulkan: load query-related functions Needed for both encoding and decoding.	1 year ago
Lynne	d4cb48dd00	hwcontext_vulkan: support PREP_MODE_DECODING in prepare_frame()	1 year ago
Lynne	6dfa29a58d	hwcontext_vulkan: add functions for video decoding	1 year ago
Lynne	571756bf2f	hwcontext_vulkan: use VK_EXT_physical_device_drm to derive DRM to Vulkan Finally, a way to directly identify a Vulkan device from a DRM device!	1 year ago
Lynne	e11fd1abdb	hwcontext_vulkan: do not require libdrm to map VAAPI devices VAAPI is sadly on the way of becoming multiplaform.	1 year ago
Lynne	f50bc930a4	hwcontext_vulkan: add support for descriptor buffers	1 year ago
Lynne	c5702abf71	hwcontext_vulkan: report nonCoherentAtomSize	1 year ago
Lynne	b0af92572f	hwcontext_vulkan: fix minor type issue in VulkanQueueCtx.buf_deps_alloc_size	1 year ago
Lynne	211c910aca	hwcontext_vulkan: rename and expand vk_pixfmt_map to append VK_NULL_FORMAT	1 year ago
Lynne	e0f2d2e702	hwcontext_vulkan: remove contiguous memory path/mode The hack was added to enable exporting of vulkan images to DRM. On Intel hardware, specifically for DRM images, all planes must be allocated next to each other, due to hardware limitation, so the hack used a single large allocation and suballocated all planes from it. By natively supporting multiplane images, the driver is what decides the layout, so exporting just works. It's a hack because it conflicted heavily with image allocation, and with the whole ecosystem in general, before multiplane images were supported, which just made it redundant. This is also the commit which broke the hwcontext hardest and prompted the entire rewrite in the first place.	1 year ago
Lynne	46a77c6496	hwcontext_vulkan: support threadsafe queue and frame operations	1 year ago

1 2 3 4 5 ...

110988 Commits (4ef5e7d47227fda15976bd70d17f325631730ff2) All Branches Search

110988 Commits (4ef5e7d47227fda15976bd70d17f325631730ff2)

All Branches