On 20/02/20 18:22, Vitaly Kuznetsov wrote: > > RFC: I looked at the code and ran some tests and nothing suspicious popped > out, however, I'm still not convinced this is a good idea to have apicv > enabled for L2 when it's disabled for L1... Also, we may prefer to merge > or re-order these two patches. I swapped the patches and queued them. The basic observation is that APICv is only about virtualizing the APIC, without any interaction with the hypervisor's APIC apart from the IPI path. So if L1 turns it on it wants L1 and L2's APICs to be completely independent, and SynIC is completely irrelevant. All that matters is again whether the IPI path works, and that is what patch 2 fixes. Paolo
