That is interesting.

Since it is built in to the Linux Kernel to setup for latency as low as 1 msec (0.001 sec) as is, and considering the 20 minute delay between Earth and Mars, it is my guess that controlled flight will not suffer from Linux latency. I imagine that autonomous behaviour liker base systems monitoring survival functions, similar to human subconscious like heartbeat, breathing, hunger, etc., and in the case of rocket engines are more likely in severe need of extremely fast response times where possibly even micro seconds matter, are what NASA was talking about. It seems to me that some proof of how low latency Linux can be is in SpaceX reusable landing methodology.

First deposit: "A titanium tube containing a rock sample is resting on the Red Planet’s surface after being placed there on Dec. 21 by NASA’s Perseverance Mars rover. Over the next two months, the rover will deposit a total of 10 tubes at the location, called “Three Forks,” building humanity’s first sample depot on another planet. The depot marks a historic early step in the Mars Sample Return campaign." https://mars.nasa.gov/news/9323/nasa...-mars-surface/

Of course, NASA knows linux is the real deal to do the job. They can't afford a blue screen on the job.

As others said, vxWorks is the typical RTOS. I am aware of one project that used RTEMS, an open-source RTOS.

I suspect that between wide ranging hardware support both old and new, low power consumption forms like ARM and RISC 5, as well as very quick onsite repurposing and reconfiguring options, Linux use will only increase at NASA and of course, SpaceX..