Andrea, I’m currently evaluating the benefits of cache coloring in jailhouse.
In my application I have a setup with two cells, the root cell for communication purposes and an inmate cell for a realtime application. The system is running on a Xilinx ZynqMP. With the arm64-zero-exits patches the realtime cell runs without hypervisor exits after startup. If io load in the root cell increases (e.g. with network traffic) I can see some impact on the performance of the realtime cell. Possible reasons might be bus congestion or cache conflicts. Since I’d like to minimize the effect, cache coloring is an obvious choice. The current set of patches for cache coloring doesn’t help a lot in my case, since only the inmate cell can be colored. In fact, the performance is worse than without cache coloring (presumably because less cache is available for the inmate). My question is this: Is there some prototype code available for root cell cache coloring? You mentioned on the list that you had done some experiments with this. I’m aware of the limitations that this may have and the concerns that Jan has voiced for this approach - however, at the present stage I’d be happy even with some hacky approach (be it in Linux, U-Boot or Jailhouse) to investigate whether cache coloring provides a benefit to my setup. Kind regards, Oliver -- You received this message because you are subscribed to the Google Groups "Jailhouse" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/jailhouse-dev/24D2D568-5334-4B88-9430-86F84E576E35%40gmx.de.
