Re: [ros-dev] [ros-diffs] [cgutman] 54418: [HALX86] - Do not allow software interrupts to preempt code running with interrupts disabled during KfLowerIrql

[Cameron Gutman]

> _disable and setting IRQL to PASSIVE_LEVEL makes no sense whatsoever. You basically hacked the HAL to support a bug in KDBG. Good job.

I gathered that what I was supporting was pretty nasty, but either way the behavior that was going on was completely wrong. I'm fairly sure there are other misuses of _disable() (and KeLowerIrql) that could result in weird behavior. I'll review them when I get some time. KDBG certainly needs to be fixed to avoid needing to manually lower IRQL.

Regards,
Cameron