Messages in this thread | | | Date | Tue, 16 Dec 2014 17:10:30 +0900 | From | Kamezawa Hiroyuki <> | Subject | Re: [PATCH 3/4] workqueue: Update workqueue's possible cpumask when a new node, coming up. |
| |
(2014/12/16 16:49), Lai Jiangshan wrote: > On 12/15/2014 07:18 PM, Kamezawa Hiroyuki wrote: >> Workqueue keeps cpu<->node relationship including all possible cpus. >> The original information was made at boot but it may change when >> a new node is added. >> >> Update information if a new node is ready with using node-hotplug callback. >> >> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> >> --- >> include/linux/memory_hotplug.h | 2 ++ >> kernel/workqueue.c | 28 +++++++++++++++++++++++++++- >> mm/memory_hotplug.c | 4 ++++ >> 3 files changed, 33 insertions(+), 1 deletion(-) >> >> diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h >> index 8f1a419..cd3cb67 100644 >> --- a/include/linux/memory_hotplug.h >> +++ b/include/linux/memory_hotplug.h >> @@ -270,4 +270,6 @@ extern void sparse_remove_one_section(struct zone *zone, struct mem_section *ms) >> extern struct page *sparse_decode_mem_map(unsigned long coded_mem_map, >> unsigned long pnum); >> >> +/* update for workqueues */ >> +void workqueue_node_register(int node); >> #endif /* __LINUX_MEMORY_HOTPLUG_H */ >> diff --git a/kernel/workqueue.c b/kernel/workqueue.c >> index 2fd0bd7..5499b76 100644 >> --- a/kernel/workqueue.c >> +++ b/kernel/workqueue.c >> @@ -266,7 +266,7 @@ struct workqueue_struct { >> static struct kmem_cache *pwq_cache; >> >> static cpumask_var_t *wq_numa_possible_cpumask; >> - /* possible CPUs of each node */ >> + /* PL: possible CPUs of each node */ >> >> static bool wq_disable_numa; >> module_param_named(disable_numa, wq_disable_numa, bool, 0444); >> @@ -4559,6 +4559,32 @@ static void restore_unbound_workers_cpumask(struct worker_pool *pool, int cpu) >> WARN_ON_ONCE(set_cpus_allowed_ptr(worker->task, >> pool->attrs->cpumask) < 0); >> } >> +#ifdef CONFIG_MEMORY_HOTPLUG >> + >> +static void reflesh_wq_possible_mask(int cpu, int node) >> +{ >> + int oldnode; >> + for_each_node_state(oldnode, N_POSSIBLE) >> + cpumask_clear_cpu(cpu, wq_numa_possible_cpumask[oldnode]); > > You need to check and update all the wq->numa_pwq_tbl[oldnode] >
you mean : if I drop patch 1/4, it will be required. Right ? I'll add that check.
Thanks, -Kame
>> + cpumask_set_cpu(cpu, wq_numa_possible_cpumask[node]); >> +} >> +/* >> + * When a cpu is physically added, cpu<->node relationship is established. >> + * We can catch the whole view when a new NODE_DATA() coming up. >> + * Update cpumask used for sched affinity of workers. >> + */ >> +void workqueue_node_register(int node) >> +{ >> + int cpu; >> + mutex_lock(&wq_pool_mutex); >> + for_each_possible_cpu(cpu) { >> + if (node == cpu_to_node(cpu)) >> + reflesh_wq_possible_mask(cpu, node); >> + } >> + mutex_unlock(&wq_pool_mutex); >> +} >> + >> +#endif >> >> /* >> * Workqueues should be brought up before normal priority CPU notifiers. >> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c >> index 1bf4807..d0c1ebb 100644 >> --- a/mm/memory_hotplug.c >> +++ b/mm/memory_hotplug.c >> @@ -1121,6 +1121,10 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start) >> * online_pages() and offline_pages(). >> */ >> reset_node_present_pages(pgdat); >> + /* >> + * Update workqueue's cpu/node affinity info. >> + */ >> + workqueue_node_register(nid); >> >> return pgdat; >> } >
| |