On 28 July 2017 at 10:18, Peter Maydell <peter.mayd...@linaro.org> wrote: > Talking to David about this on IRC, even "don't allow > migration" is a bit tricky, because you can't register > a migration-blocker if a migration is already in progress. > Two further options: > * disable this mmio-exec-in-place entirely for 2.10 as > being not completely baked yet, and fix it properly > in 2.11 > * ditto, but allow the device to have an x-enable-exec-in-place > property which turned on the not-baked feature (the > property would go away in 2.11, which is OK because > of the x- experimental prefix) > > I'm leaning towards disable-entirely at the moment...
I've sent a for-2.10 patch which takes the simple "just disable it" approach. If anybody wants to propose taking a less blunt hammer to it, they're welcome to send a patch. thanks -- PMM
