diff options
author | Dario <dariosamo@gmail.com> | 2024-05-02 15:59:29 -0300 |
---|---|---|
committer | Dario <dariosamo@gmail.com> | 2024-05-13 10:20:31 -0300 |
commit | d5789e09eb02353545124e5cb8553004b3bfc5fc (patch) | |
tree | 7bd438534002d38df72d6c1cd8418d5c3072c746 /servers/rendering/rendering_device.cpp | |
parent | c6f1f614bbab9334675026fe21f5af4951b9c890 (diff) | |
download | redot-engine-d5789e09eb02353545124e5cb8553004b3bfc5fc.tar.gz |
Add optional driver workaround to RenderingDevice for Adreno 6XX.
Co-authored-by: Clay John <claynjohn@gmail.com>
Diffstat (limited to 'servers/rendering/rendering_device.cpp')
-rw-r--r-- | servers/rendering/rendering_device.cpp | 75 |
1 files changed, 68 insertions, 7 deletions
diff --git a/servers/rendering/rendering_device.cpp b/servers/rendering/rendering_device.cpp index 31fc51efaa..6746a1dde1 100644 --- a/servers/rendering/rendering_device.cpp +++ b/servers/rendering/rendering_device.cpp @@ -4877,25 +4877,78 @@ void RenderingDevice::_end_frame() { ERR_PRINT("Found open compute list at the end of the frame, this should never happen (further compute will likely not work)."); } - draw_graph.end(frames[frame].draw_command_buffer, RENDER_GRAPH_REORDER, RENDER_GRAPH_FULL_BARRIERS); driver->command_buffer_end(frames[frame].setup_command_buffer); - driver->command_buffer_end(frames[frame].draw_command_buffer); + + // The command buffer must be copied into a stack variable as the driver workarounds can change the command buffer in use. + RDD::CommandBufferID command_buffer = frames[frame].draw_command_buffer; + draw_graph.end(RENDER_GRAPH_REORDER, RENDER_GRAPH_FULL_BARRIERS, command_buffer, frames[frame].command_buffer_pool); + driver->command_buffer_end(command_buffer); driver->end_segment(); } void RenderingDevice::_execute_frame(bool p_present) { + // Check whether this frame should present the swap chains and in which queue. const bool frame_can_present = p_present && !frames[frame].swap_chains_to_present.is_empty(); const bool separate_present_queue = main_queue != present_queue; - const VectorView<RDD::SemaphoreID> execute_draw_semaphore = frame_can_present && separate_present_queue ? frames[frame].draw_semaphore : VectorView<RDD::SemaphoreID>(); - const VectorView<RDD::SwapChainID> execute_draw_swap_chains = frame_can_present && !separate_present_queue ? frames[frame].swap_chains_to_present : VectorView<RDD::SwapChainID>(); + thread_local LocalVector<RDD::SwapChainID> swap_chains; + swap_chains.clear(); + + // Execute the setup command buffer. driver->command_queue_execute_and_present(main_queue, {}, frames[frame].setup_command_buffer, frames[frame].setup_semaphore, {}, {}); - driver->command_queue_execute_and_present(main_queue, frames[frame].setup_semaphore, frames[frame].draw_command_buffer, execute_draw_semaphore, frames[frame].draw_fence, execute_draw_swap_chains); + + // Execute command buffers and use semaphores to wait on the execution of the previous one. Normally there's only one command buffer, + // but driver workarounds can force situations where there'll be more. + uint32_t command_buffer_count = 1; + RDG::CommandBufferPool &buffer_pool = frames[frame].command_buffer_pool; + if (buffer_pool.buffers_used > 0) { + command_buffer_count += buffer_pool.buffers_used; + buffer_pool.buffers_used = 0; + } + + RDD::SemaphoreID wait_semaphore = frames[frame].setup_semaphore; + for (uint32_t i = 0; i < command_buffer_count; i++) { + RDD::CommandBufferID command_buffer; + RDD::SemaphoreID signal_semaphore; + RDD::FenceID signal_fence; + if (i > 0) { + command_buffer = buffer_pool.buffers[i - 1]; + signal_semaphore = buffer_pool.semaphores[i - 1]; + } else { + command_buffer = frames[frame].draw_command_buffer; + signal_semaphore = frames[frame].draw_semaphore; + } + + bool signal_semaphore_valid; + if (i == (command_buffer_count - 1)) { + // This is the last command buffer, it should signal the fence. + signal_fence = frames[frame].draw_fence; + signal_semaphore_valid = false; + + if (frame_can_present && separate_present_queue) { + // The semaphore is required if the frame can be presented and a separate present queue is used. + signal_semaphore_valid = true; + } else if (frame_can_present) { + // Just present the swap chains as part of the last command execution. + swap_chains = frames[frame].swap_chains_to_present; + } + } else { + // Semaphores always need to be signaled if it's not the last command buffer. + signal_semaphore_valid = true; + } + + driver->command_queue_execute_and_present(main_queue, wait_semaphore, command_buffer, signal_semaphore_valid ? signal_semaphore : VectorView<RDD::SemaphoreID>(), signal_fence, swap_chains); + + // Make the next command buffer wait on the semaphore signaled by this one. + wait_semaphore = signal_semaphore; + } + + // Indicate the fence has been signaled so the next time the frame's contents need to be used, the CPU needs to wait on the work to be completed. frames[frame].draw_fence_signaled = true; if (frame_can_present) { if (separate_present_queue) { // Issue the presentation separately if the presentation queue is different from the main queue. - driver->command_queue_execute_and_present(present_queue, frames[frame].draw_semaphore, {}, {}, {}, frames[frame].swap_chains_to_present); + driver->command_queue_execute_and_present(present_queue, wait_semaphore, {}, {}, {}, frames[frame].swap_chains_to_present); } frames[frame].swap_chains_to_present.clear(); @@ -5044,6 +5097,9 @@ Error RenderingDevice::initialize(RenderingContextDriver *p_context, DisplayServ frames[i].timestamp_cpu_result_values.resize(max_timestamp_query_elements); frames[i].timestamp_result_values.resize(max_timestamp_query_elements); frames[i].timestamp_result_count = 0; + + // Assign the main queue family and command pool to the command buffer pool. + frames[i].command_buffer_pool.pool = frames[i].command_pool; } // Start from frame count, so everything else is immediately old. @@ -5055,7 +5111,7 @@ Error RenderingDevice::initialize(RenderingContextDriver *p_context, DisplayServ driver->command_buffer_begin(frames[0].draw_command_buffer); // Create draw graph and start it initialized as well. - draw_graph.initialize(driver, frames.size(), main_queue_family, SECONDARY_COMMAND_BUFFERS_PER_FRAME); + draw_graph.initialize(driver, device, frames.size(), main_queue_family, SECONDARY_COMMAND_BUFFERS_PER_FRAME); draw_graph.begin(); for (uint32_t i = 0; i < frames.size(); i++) { @@ -5388,6 +5444,11 @@ void RenderingDevice::finalize() { driver->semaphore_free(frames[i].setup_semaphore); driver->semaphore_free(frames[i].draw_semaphore); driver->fence_free(frames[i].draw_fence); + + RDG::CommandBufferPool &buffer_pool = frames[i].command_buffer_pool; + for (uint32_t j = 0; j < buffer_pool.buffers.size(); j++) { + driver->semaphore_free(buffer_pool.semaphores[j]); + } } if (pipeline_cache_enabled) { |