On Wed, 2022-03-09 at 16:48 +0100, Paolo Bonzini wrote: > On 3/1/22 18:25, Maxim Levitsky wrote: > > > I don't like this change. It's not bad code, but it'll be confusing because it > > > implies that it's legal for svm->vmcb to be something other than svm->vmcb01.ptr > > > when this is called. > > Honestly I don't see how you had reached this conclusion. > > > > I just think that code that always works on vmcb01 > > should use it, even if it happens that vmcb == vmcb01. > > > > If you insist I can drop this patch or add WARN_ON instead, > > I just think that this way is cleaner. > > > > I do like the patch, but you should do the same in init_vmcb() and > svm_hv_init_vmcb() as well. I will do this. Thanks! Best regards, Maxim Levitsky > > Paolo >