Degrade ARM timer to monotonic raw clock.

The monotonic raw clock has the same accuracy as reading cntvct registers, but
using clock_gettime() has a measurable impact on queueing time (135 us vs
83 us).

This change is needed to enable ftrace time readings on ARM linux, which
doesn't provide any way to get raw cntvct readings, like x86-tsc on x86.

client/TracyProfiler.hpp

@@ -78,10 +78,6 @@ struct LuaZoneState
 
 using Magic = moodycamel::ConcurrentQueueDefaultTraits::index_t;
 
-#if __ARM_ARCH >= 6 && !defined TARGET_OS_IOS
-extern int64_t (*GetTimeImpl)();
-#endif
-
 
 class Profiler
 {
@@ -105,7 +101,13 @@ public:
 #  if TARGET_OS_IOS == 1
         return mach_absolute_time();
 #  elif __ARM_ARCH >= 6
-        return GetTimeImpl();
+#    ifdef CLOCK_MONOTONIC_RAW
+        struct timespec ts;
+        clock_gettime( CLOCK_MONOTONIC_RAW, &ts );
+        return int64_t( ts.tv_sec ) * 1000000000ll + int64_t( ts.tv_nsec );
+#    else
+        return std::chrono::duration_cast<std::chrono::nanoseconds>( std::chrono::high_resolution_clock::now().time_since_epoch() ).count();
+#    endif
 #  elif defined _WIN32 || defined __CYGWIN__
         unsigned int dontcare;
         const auto t = int64_t( __rdtscp( &dontcare ) );
@@ -440,8 +442,6 @@ public:
     void RequestShutdown() { m_shutdown.store( true, std::memory_order_relaxed ); m_shutdownManual.store( true, std::memory_order_relaxed ); }
     bool HasShutdownFinished() const { return m_shutdownFinished.load( std::memory_order_relaxed ); }
 
-    static bool IsTimerMonotonicRaw();
-
 private:
     enum class DequeueStatus { Success, ConnectionLost, QueueEmpty };