adlazar

> Something is weird, QEMU tells me `VMI: missing kernel built with CONFIG_KVM_INTROSPECTION` even after booting the v12 kernel with `kvm.introspection=1`. If both ([QEMU](https://github.com/KVM-VMI/qemu/commit/09bd353153176eebca420c53247e409418b29a18), kernel) are built with the same...

> @adlazar Any chance you could rebase your libvmi branch on the latest master of libvmi? I tried to do that but I'm getting: > > ``` > -o libvmi/driver/kvm/.libs/libvmi_la-kvm_events.o...

> @adlazar Any chance you could rebase your libvmi branch on the latest master of libvmi? Done. Waiting for Travis to build it :)

> but the errors are caused by the fact that the wire-protocol changed a bit @tklengyel , not quite, but now I think I've done all the necessarily changes.

> Somehow it wasn't able to start the second playbook: It fails on task/vm. I saw an error related to docker on a local run. My guess is that x86_64_defconfig...

> @adlazar LibVMI right now needs to know the max address or have a complete memory map to determine if a requested access is sane or not. I'll add support...

@tklengyel , I thought these two commits fixed the memory size issue: - [QEMU](https://github.com/KVM-VMI/qemu/commit/3b171cb719fae713d9a8832cd97249d7485e470b) sends the max gpa during handshake - [libkvmi](https://github.com/adlazar/libkvmi/commit/c2dd1d80bd970c268e57c8fa13b64ecc3f8fe920) will use that value from `kvmi_get_maximum_gfn()`

> Is the QEMU patch going to work even if the guest is booted with ovmf and there is no e820 map? I guess so - https://github.com/KVM-VMI/qemu/blob/kvmi-v12/hw/i386/pc.c#L1160

Check if ... points to a version of QEMU built with introspection patches.

@SOBXiong , you can browse the commits related to VMI starting with [aaea27d12825759b0a7649a11477eefa6744b37a]( https://github.com/KVM-VMI/qemu/commits/kvmi-v7?after=29f1a0f4c1c412e131d742a694e213ee2f0e6637+34&branch=kvmi-v7&qualified_name=refs%2Fheads%2Fkvmi-v7), the first commit on top of v4.2.1