16:51 <chrisl> I'm trying to debug the "bootloop" that happens on m1n1 booting macOS 14.0+. Using the Central Scrutinizer I was able to quickly determine that the m1n1 guest becomes unresponsive relatively early in boot at the end of mmu_init when enabling the MMU. But I've been stuck at that point for a while now, and I was wondering if anyone had any good tips for debugging *why* enabling the MMU is causing problems.

16:52 <chrisl> I'm slowly learning more things (for instance, I didn't think SCTLR_EL1 should matter since we're running in EL2 at that point until I read about HCR.E2H), but I tried instrumenting all the exception vectors thinking there might be a page fault, but I'm not getting any serial output, so if a exception is being triggered, it's not being handled the way I expect it to.

16:53 <chrisl> Without the MMU enabled, I can't get the actual serial port up to the host to be able to issue commands, so my debugging is currently at the level of rebuilding m1n1, booting into 1TR, reconfiguring and then trying again for every little change

16:54 <chrisl> I'm guessing that iBoot starting configuring the system differently at the 14.0 boundary, and something about the current state is causing the hang when enabling the MMU (since it worked in some 14.0 betas, but not as of full release), so if there's a way to see what's different in the iBoot, or system config at that point that might point out the differences