Checking out -current via CVS is just kinda dreadful. I'm not thrilled on building this on a Pi either, but it's good to at least have a not entirely broken example either. I do wonder if it would be faster to check out elsewhere and rsync
Anyway, back home. Decided to get back to the OpenBSD/arm64 issues. I threw it on a Pi, and its not spinning which means its at least not a general platform issue.
I think the next step is to build the OpenBSD kernel singleprocessing and see if the gprof output gives more clues.
It's spending a lot of time dealing with VM related stuff, but after staring at the code for awhile, nothing is jumping out. Single core performance was a lot better though whcih makes me suspect this is a SMP issue.
@onekopaka Well I saw that it was doing the same TOP behavior only showing half the processors, which didn't quite register as why it was weird. But that's why I want to profile it