On Tue, 15 Jan 2013 19:43:40 -0500
Christoffer Dall <cdall@xxxxxxxxxxxxxxx> wrote:
> @@ -201,9 +222,13 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)
> {
> long dirty_count = kvm->tlbs_dirty;
>
> - smp_mb();
> + smp_mb(); /* TODO: Someone should explain this! */
> if (make_all_cpus_request(kvm, KVM_REQ_TLB_FLUSH))
> ++kvm->stat.remote_tlb_flush;
> + /*
> + * TODO: Someone should explain this, why is it a cmpxchg, what
> + * happens if the dirty is different from dirty_count?
> + */
> cmpxchg(&kvm->tlbs_dirty, dirty_count, 0);
> }
If it's different from the old value, that means another
attempt which forces us to care about TLBs has been done?
So I think making tlbs_dirty 0 is safe only when it's not
changed during executing this function.
Takuya
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
