Make the device clock members atomic

Even though they're protected by a SeqLock of sorts, it's still UB to read and
write non-atomic vars from different threads. It's fine to do relaxed reads and
writes given the lock though, to help alleviate the cost.

1 files changed, 3 insertions, 0 deletions

diff --git a/core/device.cpp b/core/device.cpp
index 2766c5e4..a5edf63c 100644
--- a/core/device.cpp
+++ b/core/device.cpp
@@ -9,6 +9,9 @@
 #include "mastering.h"
 
 
+static_assert(std::atomic<std::chrono::nanoseconds>::is_always_lock_free);
+
+
 al::FlexArray<ContextBase*> DeviceBase::sEmptyContextArray{0u};