Frozen Netfilter Queues for processing IP packets from userspace
Issue
When a multi-threaded application starts up, and several threads invoke nfq_open_nfnl() at the exactly same time, its possible that one thread ends up with another threads fq_handle in its private per-thread data. When this happens, one queue is seemingly "forgotten" and no packets will be processed because several threads erroneously process the same queue.
Environment
- rhel-7
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.