Add perfetto tracepoints to async record

app/src/main/cpp/skyline/gpu/interconnect/command_executor.cpp

@@ -37,6 +37,7 @@ namespace skyline::gpu::interconnect {
     }
 
     void CommandRecordThread::ProcessSlot(Slot *slot) {
+        TRACE_EVENT_FMT("gpu", "ProcessSlot: 0x{:X}, execution: {}", slot, slot->executionNumber);
         auto &gpu{*state.gpu};
 
         vk::RenderPass lRenderPass;
@@ -129,6 +130,7 @@ namespace skyline::gpu::interconnect {
 
         slot = recordThread.AcquireSlot();
         cycle = slot->Reset(gpu);
+        slot->executionNumber = executionNumber;
         allocator = &slot->allocator;
     }
 
@@ -411,12 +413,12 @@ namespace skyline::gpu::interconnect {
         for (const auto &callback : flushCallbacks)
             callback();
 
+        executionNumber++;
+
         if (!slot->nodes.empty()) {
             TRACE_EVENT("gpu", "CommandExecutor::Submit");
             SubmitInternal();
         }
         ResetInternal();
-
-        executionNumber++;
     }
 }

app/src/main/cpp/skyline/gpu/interconnect/command_executor.h

@@ -25,6 +25,7 @@ namespace skyline::gpu::interconnect {
             std::shared_ptr<FenceCycle> cycle;
             boost::container::stable_vector<node::NodeVariant> nodes;
             LinearAllocatorState<> allocator;
+            u32 executionNumber;
 
             Slot(GPU &gpu);
 
@@ -151,7 +152,7 @@ namespace skyline::gpu::interconnect {
         std::shared_ptr<FenceCycle> cycle; //!< The fence cycle that this command executor uses to wait for the GPU to finish executing commands
         LinearAllocatorState<> *allocator;
         ContextTag tag; //!< The tag associated with this command executor, any tagged resource locking must utilize this tag
-        size_t executionNumber{};
+        u32 executionNumber{};
 
         CommandExecutor(const DeviceState &state);