From 36f20ae3b2975e44b6ef17e453ae06a289e9a122 Mon Sep 17 00:00:00 2001 From: Kyle Walker Date: Thu, 30 Jun 2016 15:12:18 -0400 Subject: manager: Only invoke a single sigchld per unit within a cleanup cycle By default, each iteration of manager_dispatch_sigchld() results in a unit level sigchld event being invoked. For scope units, this results in a scope_sigchld_event() which can seemingly stall for workloads that have a large number of PIDs within the scope. The stall exhibits itself as a SIG_0 being initiated for each u->pids entry as a result of pid_is_unwaited(). v2: This patch resolves this condition by only paying to cost of a sigchld in the underlying scope unit once per sigchld iteration. A new "sigchldgen" member resides within the Unit struct. The Manager is incremented via the sd event loop, accessed via sd_event_get_iteration, and the Unit member is set to the same value as the manager each time that a sigchld event is invoked. If the Manager iteration value and Unit member match, the sigchld event is not invoked for that iteration. --- src/core/unit.h | 3 +++ 1 file changed, 3 insertions(+) (limited to 'src/core/unit.h') diff --git a/src/core/unit.h b/src/core/unit.h index 08a927962d..c41011ed9d 100644 --- a/src/core/unit.h +++ b/src/core/unit.h @@ -162,6 +162,9 @@ struct Unit { * process SIGCHLD for */ Set *pids; + /* Used in sigchld event invocation to avoid repeat events being invoked */ + uint64_t sigchldgen; + /* Used during GC sweeps */ unsigned gc_marker; -- cgit v1.2.3-54-g00ecf