From: Zhang, X. <xia...@in...> - 2007-11-09 07:25:40
|
Hollis Blanchard wrote: >> - kvm_x86_ops->vcpu_free(vcpu); >> + kvm_arch_vcpu_free(vcpu); >> return r; >> } >=20 > Have a look at the patch I posted on Wednesday: "[PATCH 2 of 2] RFC: > Create kvm_arch_vcpu_create()". kvm_arch_vcpu_create() will actually > encompass more logic from kvm_vm_ioctl_create_vcpu(), and once you do > that, kvm_arch_vcpu_reset() doesn't need to exist (which will also > make Carsten happy). Hi, Hollis Thanks for your response. If we need to move out the code related to mmu, sure to keep more logic in kvm_arch_create_vcpu. But I don't know the decision about the mmu part putting in common or arch. As Avi said before, maybe we need to keep mmu interfaces in common for future possible memory hotplug of guest and so on. =20 Best Wishes Xiantao |