Hide rpmalloc init behind thread local boolean.

client/TracyAlloc.cpp

@@ -0,0 +1,35 @@
+#ifdef TRACY_ENABLE
+
+#include <atomic>
+
+#include "../common/TracyAlloc.hpp"
+#include "../common/TracyYield.hpp"
+
+namespace tracy
+{
+
+extern std::atomic<int> RpInitDone;
+extern std::atomic<int> RpInitLock;
+
+TRACY_API void InitRpmallocPlumbing()
+{
+    const auto done = RpInitDone.load( std::memory_order_acquire );
+    if( !done )
+    {
+        int expected = 0;
+        while( !RpInitLock.compare_exchange_weak( expected, 1, std::memory_order_release, std::memory_order_relaxed ) ) { expected = 0; YieldThread(); }
+        const auto done = RpInitDone.load( std::memory_order_acquire );
+        if( !done )
+        {
+            rpmalloc_initialize();
+            RpInitDone.store( 1, std::memory_order_release );
+        }
+        RpInitLock.store( 0, std::memory_order_release );
+    }
+    rpmalloc_thread_initialize();
+    RpThreadInitDone = true;
+}
+
+}
+
+#endif

client/TracyProfiler.cpp

@@ -977,6 +977,7 @@ TRACY_API void ShutdownProfiler()
 #  else
 std::atomic<int> RpInitDone { 0 };
 std::atomic<int> RpInitLock { 0 };
+thread_local bool RpThreadInitDone = false;
 static std::atomic<int> profilerDataLock { 0 };
 static std::atomic<ProfilerData*> profilerData { nullptr };
 
@@ -1097,6 +1098,7 @@ thread_local ThreadHandleWrapper init_order(104) s_threadHandle { detail::GetThr
 static InitTimeWrapper init_order(101) s_initTime { SetupHwTimer() };
 std::atomic<int> init_order(102) RpInitDone( 0 );
 std::atomic<int> init_order(102) RpInitLock( 0 );
+thread_local bool RpThreadInitDone = false;
 moodycamel::ConcurrentQueue<QueueItem> init_order(103) s_queue( QueuePrealloc );
 std::atomic<uint32_t> init_order(104) s_lockCounter( 0 );
 std::atomic<uint8_t> init_order(104) s_gpuCtxCounter( 0 );