On Tuesday, January 24, 2023 7:25 AM, Sean Christopherson wrote: > > - kfree(bus); > > - return new_bus ? 0 : -ENOMEM; > > + kvm_iodevice_destructor(dev); > > + return 0; > > Unless I'm misreading things, this path leaks "bus". Right, should keep the kfree above. > Given that that intent is to send the fix for stable, that this is as much a > cleanup as it is a bug fix, and that it's not super trivial, I'm inclined to queue > my patch and then land this on top as cleanup. Sounds good to me.