diff options
Diffstat (limited to '')
-rw-r--r-- | src/video_core/renderer_vulkan/vk_stream_buffer.cpp | 142 |
1 files changed, 100 insertions, 42 deletions
diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp index 62f1427f5..d48d3b44c 100644 --- a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp +++ b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp @@ -3,86 +3,144 @@ // Refer to the license.txt file included. #include <algorithm> -#include <memory> #include <optional> +#include <tuple> #include <vector> +#include "common/alignment.h" #include "common/assert.h" #include "video_core/renderer_vulkan/declarations.h" #include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_stream_buffer.h" namespace Vulkan { +namespace { + constexpr u64 WATCHES_INITIAL_RESERVE = 0x4000; constexpr u64 WATCHES_RESERVE_CHUNK = 0x1000; -VKStreamBuffer::VKStreamBuffer(const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler, u64 size, vk::BufferUsageFlags usage, - vk::AccessFlags access, vk::PipelineStageFlags pipeline_stage) - : device{device}, scheduler{scheduler}, buffer_size{size}, access{access}, pipeline_stage{ - pipeline_stage} { - CreateBuffers(memory_manager, usage); - ReserveWatches(WATCHES_INITIAL_RESERVE); +constexpr u64 STREAM_BUFFER_SIZE = 256 * 1024 * 1024; + +std::optional<u32> FindMemoryType(const VKDevice& device, u32 filter, + vk::MemoryPropertyFlags wanted) { + const auto properties = device.GetPhysical().getMemoryProperties(device.GetDispatchLoader()); + for (u32 i = 0; i < properties.memoryTypeCount; i++) { + if (!(filter & (1 << i))) { + continue; + } + if ((properties.memoryTypes[i].propertyFlags & wanted) == wanted) { + return i; + } + } + return {}; +} + +} // Anonymous namespace + +VKStreamBuffer::VKStreamBuffer(const VKDevice& device, VKScheduler& scheduler, + vk::BufferUsageFlags usage) + : device{device}, scheduler{scheduler} { + CreateBuffers(usage); + ReserveWatches(current_watches, WATCHES_INITIAL_RESERVE); + ReserveWatches(previous_watches, WATCHES_INITIAL_RESERVE); } VKStreamBuffer::~VKStreamBuffer() = default; -std::tuple<u8*, u64, bool> VKStreamBuffer::Reserve(u64 size) { - ASSERT(size <= buffer_size); +std::tuple<u8*, u64, bool> VKStreamBuffer::Map(u64 size, u64 alignment) { + ASSERT(size <= STREAM_BUFFER_SIZE); mapped_size = size; - if (offset + size > buffer_size) { - // The buffer would overflow, save the amount of used buffers, signal an invalidation and - // reset the state. - invalidation_mark = used_watches; - used_watches = 0; + if (alignment > 0) { + offset = Common::AlignUp(offset, alignment); + } + + WaitPendingOperations(offset); + + bool invalidated = false; + if (offset + size > STREAM_BUFFER_SIZE) { + // The buffer would overflow, save the amount of used watches and reset the state. + invalidation_mark = current_watch_cursor; + current_watch_cursor = 0; offset = 0; + + // Swap watches and reset waiting cursors. + std::swap(previous_watches, current_watches); + wait_cursor = 0; + wait_bound = 0; + + // Ensure that we don't wait for uncommitted fences. + scheduler.Flush(); + + invalidated = true; } - return {mapped_pointer + offset, offset, invalidation_mark.has_value()}; + const auto dev = device.GetLogical(); + const auto& dld = device.GetDispatchLoader(); + const auto pointer = reinterpret_cast<u8*>(dev.mapMemory(*memory, offset, size, {}, dld)); + return {pointer, offset, invalidated}; } -void VKStreamBuffer::Send(u64 size) { +void VKStreamBuffer::Unmap(u64 size) { ASSERT_MSG(size <= mapped_size, "Reserved size is too small"); - if (invalidation_mark) { - // TODO(Rodrigo): Find a better way to invalidate than waiting for all watches to finish. - scheduler.Flush(); - std::for_each(watches.begin(), watches.begin() + *invalidation_mark, - [&](auto& resource) { resource->Wait(); }); - invalidation_mark = std::nullopt; - } + const auto dev = device.GetLogical(); + dev.unmapMemory(*memory, device.GetDispatchLoader()); + + offset += size; - if (used_watches + 1 >= watches.size()) { + if (current_watch_cursor + 1 >= current_watches.size()) { // Ensure that there are enough watches. - ReserveWatches(WATCHES_RESERVE_CHUNK); + ReserveWatches(current_watches, WATCHES_RESERVE_CHUNK); } - // Add a watch for this allocation. - watches[used_watches++]->Watch(scheduler.GetFence()); - - offset += size; + auto& watch = current_watches[current_watch_cursor++]; + watch.upper_bound = offset; + watch.fence.Watch(scheduler.GetFence()); } -void VKStreamBuffer::CreateBuffers(VKMemoryManager& memory_manager, vk::BufferUsageFlags usage) { - const vk::BufferCreateInfo buffer_ci({}, buffer_size, usage, vk::SharingMode::eExclusive, 0, - nullptr); - +void VKStreamBuffer::CreateBuffers(vk::BufferUsageFlags usage) { + const vk::BufferCreateInfo buffer_ci({}, STREAM_BUFFER_SIZE, usage, vk::SharingMode::eExclusive, + 0, nullptr); const auto dev = device.GetLogical(); const auto& dld = device.GetDispatchLoader(); buffer = dev.createBufferUnique(buffer_ci, nullptr, dld); - commit = memory_manager.Commit(*buffer, true); - mapped_pointer = commit->GetData(); + + const auto requirements = dev.getBufferMemoryRequirements(*buffer, dld); + // Prefer device local host visible allocations (this should hit AMD's pinned memory). + auto type = FindMemoryType(device, requirements.memoryTypeBits, + vk::MemoryPropertyFlagBits::eHostVisible | + vk::MemoryPropertyFlagBits::eHostCoherent | + vk::MemoryPropertyFlagBits::eDeviceLocal); + if (!type) { + // Otherwise search for a host visible allocation. + type = FindMemoryType(device, requirements.memoryTypeBits, + vk::MemoryPropertyFlagBits::eHostVisible | + vk::MemoryPropertyFlagBits::eHostCoherent); + ASSERT_MSG(type, "No host visible and coherent memory type found"); + } + const vk::MemoryAllocateInfo alloc_ci(requirements.size, *type); + memory = dev.allocateMemoryUnique(alloc_ci, nullptr, dld); + + dev.bindBufferMemory(*buffer, *memory, 0, dld); } -void VKStreamBuffer::ReserveWatches(std::size_t grow_size) { - const std::size_t previous_size = watches.size(); - watches.resize(previous_size + grow_size); - std::generate(watches.begin() + previous_size, watches.end(), - []() { return std::make_unique<VKFenceWatch>(); }); +void VKStreamBuffer::ReserveWatches(std::vector<Watch>& watches, std::size_t grow_size) { + watches.resize(watches.size() + grow_size); +} + +void VKStreamBuffer::WaitPendingOperations(u64 requested_upper_bound) { + if (!invalidation_mark) { + return; + } + while (requested_upper_bound < wait_bound && wait_cursor < *invalidation_mark) { + auto& watch = previous_watches[wait_cursor]; + wait_bound = watch.upper_bound; + watch.fence.Wait(); + ++wait_cursor; + } } } // namespace Vulkan |