On Wed Feb 12, 2025 at 9:11 AM CET, Maurizio Lombardi wrote: > Hello, could you try this patch? > > Concurrent calls to try_recv() should already be protected by > sock_lock. > > + mutex_lock(&queue->send_mutex); > nvme_tcp_try_recv(queue); > + r = queue->nr_cqe; > + mutex_unlock(&queue->send_mutex); Well, reading nr_cqe like this is still racy, but should be a minor issue and not hard to fix. Maurizio