On Wed 03-02-16 13:54:25, Tejun Heo wrote:
> When looking up the pool_workqueue to use for an unbound workqueue,
> workqueue assumes that the target CPU is always bound to a valid NUMA
> node. However, currently, when a CPU goes offline, the mapping is
> destroyed and cpu_to_node() returns NUMA_NO_NODE. This has always
> been broken but hasn't triggered until recently.
>
> After 874bbfe600a6 ("workqueue: make sure delayed work run in local
> cpu"), workqueue forcifully assigns the local CPU for delayed work
> items without explicit target CPU to fix a different issue. This
> widens the window where CPU can go offline while a delayed work item
> is pending causing delayed work items dispatched with target CPU set
> to an already offlined CPU. The resulting NUMA_NO_NODE mapping makes
> workqueue try to queue the work item on a NULL pool_workqueue and thus
> crash.
>
> Fix it by mapping NUMA_NO_NODE to the default pool_workqueue from
> unbound_pwq_by_node(). This is a temporary workaround. The long term
> solution is keeping CPU -> NODE mapping stable across CPU off/online
> cycles which is in the works.
>
> Signed-off-by: Tejun Heo <[email protected]>
> Reported-by: Mike Galbraith <[email protected]>
> Cc: Tang Chen <[email protected]>
> Cc: Rafael J. Wysocki <[email protected]>
> Cc: Len Brown <[email protected]>
> Cc: [email protected] # v4.3+
> Fixes: 874bbfe600a6 ("workqueue: make sure delayed work run in local cpu")
> Link: http://lkml.kernel.org/g/[email protected]
> Link:
> http://lkml.kernel.org/g/[email protected]
Reviewed-by: Michal Hocko <[email protected]>
> ---
> kernel/workqueue.c | 10 ++++++++++
> 1 file changed, 10 insertions(+)
>
> diff --git a/kernel/workqueue.c b/kernel/workqueue.c
> index 61a0264..f748eab 100644
> --- a/kernel/workqueue.c
> +++ b/kernel/workqueue.c
> @@ -570,6 +570,16 @@ static struct pool_workqueue *unbound_pwq_by_node(struct
> workqueue_struct *wq,
> int node)
> {
> assert_rcu_or_wq_mutex_or_pool_mutex(wq);
> +
> + /*
> + * XXX: @node can be NUMA_NO_NODE if CPU goes offline while a
> + * delayed item is pending. The plan is to keep CPU -> NODE
> + * mapping valid and stable across CPU on/offlines. Once that
> + * happens, this workaround can be removed.
> + */
> + if (unlikely(node == NUMA_NO_NODE))
> + return wq->dfl_pwq;
> +
> return rcu_dereference_raw(wq->numa_pwq_tbl[node]);
> }
>
--
Michal Hocko
SUSE Labs
