Hi Minchan,
Thanks for comments and will send out a separate patch for
readability soon based on your version.
Thanks!
Gerry
On 2012-6-14 13:31, Minchan Kim wrote:
> Hi,
>
> On 06/14/2012 12:44 PM, Jiang Liu wrote:
>
>> Function kswapd_stop() will be called to destroy the kswapd work thread
>> when all memory of a NUMA node has been offlined. But kswapd_stop() only
>> terminates the work thread without resetting NODE_DATA(nid)->kswapd to NULL.
>> The stale pointer will prevent kswapd_run() from creating a new work thread
>> when adding memory to the memory-less NUMA node again. Eventually the stale
>> pointer may cause invalid memory access.
>>
>> Signed-off-by: Xishi Qiu <qiuxishi@xxxxxxxxxx>
>> Signed-off-by: Jiang Liu <liuj97@xxxxxxxxx>
>
>
> Reviewed-by: Minchan Kim <minchan@xxxxxxxxxx>
>
> Nitpick:
>
> I saw kswapd_run and doubt why following line is there.
>
> if (pgdat->kswapd)
> return 0;
>
> As looking thorough hotplug, I realized one can hotplug pages which are within different zones but same node.
> Because kswapd live in per-node, that code is for checking kswapd already run. Right?
Yes, I think so. We could also add new memory pages to existing zones too.
>
> IMHO, better readable code is following as
>
> diff --git a/include/linux/swap.h b/include/linux/swap.h
> index b967eda..9425c0e 100644
> --- a/include/linux/swap.h
> +++ b/include/linux/swap.h
> @@ -299,6 +299,7 @@ static inline void scan_unevictable_unregister_node(struct node *node)
> }
> #endif
>
> +extern bool is_kswapd_running(int nid);
> extern int kswapd_run(int nid);
> extern void kswapd_stop(int nid);
> #ifdef CONFIG_CGROUP_MEM_RES_CTLR
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index 0d7e3ec..60f9155 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -522,7 +522,8 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages)
> init_per_zone_wmark_min();
>
> if (onlined_pages) {
> - kswapd_run(zone_to_nid(zone));
> + if (!is_kswapd_running(zone_to_nid(zone))
> + kswapd_run(zone_to_nid(zone));
> node_set_state(zone_to_nid(zone), N_HIGH_MEMORY);
> }
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index eeb3bc9..f331904 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -2932,6 +2932,14 @@ static int __devinit cpu_callback(struct notifier_block *nfb,
> return NOTIFY_OK;
> }
>
> +bool is_kswapd_running(int nid)
> +{
> + pg_data_t *pgdat = NODE_DATA(nid);
> + if (pgdat->kswapd)
> + return true;
> + return false;
> +}
> +
> /*
> * This kswapd start function will be called by init and node-hot-add.
> * On node-hot-add, kswapd will moved to proper cpus if cpus are hot-added.
> @@ -2941,9 +2949,6 @@ int kswapd_run(int nid)
> pg_data_t *pgdat = NODE_DATA(nid);
> int ret = 0;
>
> - if (pgdat->kswapd)
> - return 0;
> -
> pgdat->kswapd = kthread_run(kswapd, pgdat, "kswapd%d", nid);
> if (IS_ERR(pgdat->kswapd)) {
> /* failure at boot is fatal */
>
> Anyway, it's a preference and trivial but I hope you fix that, too if you don't mind
> Of course, my nitpick shouldn't prevent merging your good fix.
> If you mind it, I don't care of it. :)
>
> Thanks.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>