Hi, There is no "write_lock" in RCU per se. The update side uses a synchronize_rcu() instead.
Yes, there can be interference between multiple shared resources using the same liburcu flavor. Very long read-side critical sections can delay completion of synchronize_rcu. I've done prototypes of liburcu flavors that have a "domain" structure, thus allowing each liburcu users to have their own grace period domain, but never had time to complete it and get it upstream. Thanks, Mathieu ----- On Apr 9, 2018, at 2:32 AM, 志昌 余 <yuzhichang_...@hotmail.com> wrote: > According to [ > https://github.com/urcu/userspace-rcu/blob/master/doc/rcu-api.md > | https://github.com/urcu/userspace-rcu/blob/master/doc/rcu-api.md ] , none > librcu API take the resource address. > If there're multiple independent shared resource, seems there's i nterference > between read_lock of resource A and write_lock of resource B. > _______________________________________________ > lttng-dev mailing list > email@example.com > https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev -- Mathieu Desnoyers EfficiOS Inc. http://www.efficios.com
_______________________________________________ lttng-dev mailing list firstname.lastname@example.org https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev