Do not execute exception handlers with disabled interrupts

[kib]

Summary

We should not call vm_fault(), or send a signal, with interrupts disabled. MI kernel code is not prepared for such environment, not to mention that this increases the system latency, since code appears to be executing as being under spinlock.

The FAR register for data aborts is read before the interrupts are enabled, to avoid its corruption due to nested exception or context switch.

[andrew]

sys/arm64/arm64/trap.c

79 ↗

(On Diff #8762)

We should probably move print_registers to the top of the file.

[kib]

WRT print_registers, I would like to eventually add analog of x86 machdep.uprintf_signal and enhance the panic with more MD data. I do not see much sense in increasing churn right now, most of other kernel code authors do not bother with reordering functions in file to ensure local visibility without forward declaration.