On Wed, Feb 07, 2024 at 07:39:50AM -0800, Sean Christopherson wrote: > On Thu, Nov 09, 2023, Anish Moorthy wrote: > > On Thu, Nov 9, 2023 at 1:03 PM Anish Moorthy <amoorthy@xxxxxxxxxx> wrote: > > > > > > TODO: Changelog -- and possibly just merge into the "god" arm commit? > > > > *Facepalm* > > > > Well as you can tell, I wasn't sure if there was anything to actually > > put in the long-form log. Lmk if you have suggestions > > I think the right way to organize things is to have this chunk: > > diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c > index b1e5e42bdeb4..bc978260d2be 100644 > --- a/arch/x86/kvm/mmu/mmu.c > +++ b/arch/x86/kvm/mmu/mmu.c > @@ -3309,6 +3309,10 @@ static int kvm_handle_error_pfn(struct kvm_vcpu *vcpu, struct kvm_page_fault *fa > return RET_PF_RETRY; > } > > + WARN_ON_ONCE(fault->goal_level != PG_LEVEL_4K); > + > + kvm_prepare_memory_fault_exit(vcpu, gfn_to_gpa(fault->gfn), PAGE_SIZE, > + fault->write, fault->exec, fault->is_private); > return -EFAULT; > } Err.. This is the arm64 patch. x86 already advertises KVM_CAP_MEMORY_FAULT_INFO. The rest of the advertisement happens over in the arch-neutral code when the arch selects CONFIG_HAVE_KVM_EXIT_ON_MISSING. Having said that... > be part of this patch. Because otherwise, advertising KVM_CAP_MEMORY_FAULT_INFO > is a lie. Userspace can't catch KVM in the lie, but that doesn't make it right. > > That should in turn make it easier to write a useful changelog. The feedback still stands. The capability needs to be squashed into the patch that actually introduces the functionality. -- Thanks, Oliver
