On Wed, 2014-06-11 at 12:47 +0200, Paolo Bonzini wrote: > Ok, I can buy removing the support for CPU != 0. But still, the > overengineering remains. > > If Alexey needs to trigger the NMI on all CPUs, simply moving the new > method from CPU to Machine, and blindly using first_cpu in s390 and x86 > makes the most sense.
I can imagine a future where we would want to programatically trigger the sreset (the NMI is basically a soft reset for us) on a specific target CPU so it does make some sense to have a per CPU method and have the monitor "nmi" command call it on all of them. But it's not a huge deal if you decide to just make it as machine for now. This is more than just a "debug" feature however, it is generally used to trigger kdump on crashed partitions (this feature is called "dump" in pHyp). Cheers, Ben.