On Mon, Jan 27, 2025 at 6:27 PM Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
> On Sat, Jan 25, 2025 at 1:44 AM Sean Christopherson <seanjc@xxxxxxxxxx> wrote:
> > I like the special casing, it makes the oddballs stand out, which in turn (hopefully)
> > makes developers pause and take note. I.e. the SRCU walkers are all normal readers,
> > the set_nx_huge_pages() "never" path is a write in disguise, and
> > kvm_hyperv_tsc_notifier() is a very special snowflake.
>
> Likewise, kvm_hyperv_tsc_notifier()'s requirement does deserve a comment,
> but its specialness is self-inflicted pain due to using (S)RCU even when
> it's not the most appropriate synchronization mechanism.
... in fact, you could have a KVM_CREATE_VCPU and KVM_RUN after this
point:
mutex_lock(&kvm_lock);
list_for_each_entry(kvm, &vm_list, vm_list)
kvm_make_mclock_inprogress_request(kvm);
because kvm_lock is not enough to ensure that all vCPUs got the
KVM_REQ_MCLOCK_INPROGRESS memo. So kvm_hyperv_tsc_notifier()'s
complications go beyond kvm_lock and the choice to use SRCU or not.
Paolo
