Hacker News new | past | comments | ask | show | jobs | submit login

The infection is the other way around here, the producer thread is slowed down due to a low priority of the consumer.

TLDR for article: the consumer has a lower priority and pauses on the context switch of the mutex lock, but doesn't get rescheduled for a while. The producer then has to wait for the consumer to get rescheduled and unlock the mutex before it can enqueue something.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: