On 12/3/24 08:55, Gustavo Romero wrote:
It's not "really" using kvm acceleration right?
It's using kvm but the kvm module in the host kernel is actually running
fully emulated in QEMU system mode (-accel tcg is being passed to it).
So I understand it's "all TCG" in the end and no HW acceleration is ever used
in the tests.
Ehh.. yes and no.
Yes, the system as a whole is running on a host under TCG.
No, the nested guest is itself using KVM interfaces to the guest kernel. In this case,
it's what we must to do in order to exercise the FEAT_RME code within TCG.
But contrast this with another scenario in which the nested guest is *also* running under
TCG. Double jit translation is, as you might imagine, a bit slow. But it can be useful
for smoke testing code running in the outermost guest.
r~
