it means that the OS can dynamically change the QoS of the threads
even if we are explicitly setting it.
Well, it depends on your perspective. There’s a QoS value that you control and priority value that’s initially derived from that QoS but can differ due to priority boosting. If you have a high QoS thread waiting on a resource held by a low QoS thread, the system will boost the priority of the low QoS thread to minimise the impact of this priority inversion.
Also, the main thread has QoS value USER_INTERACTIVE
That’s not always the case. Check out the doc comments for qos_class_main.
is it the case that its child threads will inherit the QoS value from
the parent thread
No. Consider the code below. I put it in a command-line tool project and the resulting tool prints:
main: userInteractive
child: default
Share and Enjoy
—
Quinn “The Eskimo!” @ Developer Technical Support @ Apple
let myEmail = "eskimo" + "1" + "@" + "apple.com"
import Darwin
extension qos_class_t {
var name: String {
switch self {
case QOS_CLASS_USER_INTERACTIVE: return "userInteractive"
case QOS_CLASS_USER_INITIATED: return "userInitiated"
case QOS_CLASS_DEFAULT: return "default"
case QOS_CLASS_UTILITY: return "utility"
case QOS_CLASS_BACKGROUND: return "background"
case QOS_CLASS_UNSPECIFIED: return "unspecified"
default: return "\(self.rawValue)?"
}
}
}
func main() {
print("main:", qos_class_self().name)
var t: pthread_t? = nil
let err = pthread_create(&t, nil, { _ in
print("child:", qos_class_self().name)
return nil
}, nil)
assert(err == 0)
let err2 = pthread_join(t!, nil)
assert(err2 == 0)
}
main()
