Hi Andrew, when I sent an earlier version of the posix mqueue namespaces patchset you called me on setting init_ipc_ns.count to 2. The following patch changes that to 1. This is safe because the 1-count is for swapper, who cannot unshare the namespace. Having init do unshare(CLONE_NEWIPC) only drops the count to 1. No pre-init resources other than tasks pin (and therefore can decrease the refcount on) an ipc_ns.
Signed-off-by: Serge E. Hallyn <[email protected]> --- ipc/msgutil.c | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/ipc/msgutil.c b/ipc/msgutil.c index e5d6da1..f095ee2 100644 --- a/ipc/msgutil.c +++ b/ipc/msgutil.c @@ -26,8 +26,7 @@ DEFINE_SPINLOCK(mq_lock); * and not CONFIG_IPC_NS. */ struct ipc_namespace init_ipc_ns = { - /* It's not for this patch to change, but should this be 1? */ - .count = ATOMIC_INIT(2), + .count = ATOMIC_INIT(1), #ifdef CONFIG_POSIX_MQUEUE .mq_queues_max = DFLT_QUEUESMAX, .mq_msg_max = DFLT_MSGMAX, -- 1.5.4.3 _______________________________________________ Containers mailing list [email protected] https://lists.linux-foundation.org/mailman/listinfo/containers _______________________________________________ Devel mailing list [email protected] https://openvz.org/mailman/listinfo/devel
