2
$\begingroup$

enter image description here

The above answer taken from the book https://www.amazon.com/Operating-Systems-Sibsankar-Haldar/dp/8131730220 says starvation freedom may not imply progress as processes can be deadlocked.

I think that starvation freedom implies deadlock freedom so there is no chance of deadlock and progress is also satisfied.

Is the answer in the picture correct ?

$\endgroup$
-1
$\begingroup$

It depends on how exactly you define "starvation freedom" and "liveness."

I believe the book is talking about starvation freedom from the perspective of the OS scheduler, which usually does not know about locks managed by user processes. So, the book is partially right (in some context).

Your definition is more broad and is what someone working in distributed systems would think (see https://en.wikipedia.org/wiki/Liveness. Note the important assumption "All processes are assumed to correctly use the mutex...").

However, I'd think that "liveness implies starvation freedom," but there's no justification in the book for the first statement in the answer.

$\endgroup$
  • $\begingroup$ Even from OS perspective starvation freedom implies deadlock freedom. Also, first statement is correct (progress doesn't imply starvation freedom) and I found out an example for that. I am not sure about the second statement. $\endgroup$ – Zephyr Jan 11 '18 at 19:44
  • $\begingroup$ Could you define starvation freedom so we're on the same page? $\endgroup$ – Vimal Jan 12 '18 at 20:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.