Fix some VTS failures using HWC3
After running VTS on a different device, I ran into some consistent failures which ended up being race conditions. One was simply failing to hold the main lock for executeCommands
, but the other one was a race between destruction of the HwcDisplay and its associated VSyncWorker thread.
I see a short sleep used to mitigate this issue elsewhere, so I took the same approach. I'd like to follow up on this to see if this can be handled more robustly without relying on sleeps.