[RFC PATCH 2/12] memory-hogplug : check memory offline in offline_pages

[KOSAKI Motohiro]

On Wed, Jun 27, 2012 at 1:44 AM, Yasuaki Ishimatsu
<isimatu.yasuaki at jp.fujitsu.com> wrote:
> When offline_pages() is called to offlined memory, the function fails since
> all memory has been offlined. In this case, the function should succeed.
> The patch adds the check function into offline_pages().

I don't understand your point. I think following misoperation should
fail. Otherwise
administrator have no way to know their fault.

$ echo offline > memoryN/state
$ echo offline > memoryN/state

In general, we don't like to ignore an error except the standard require it.

>
> CC: Len Brown <len.brown at intel.com>
> CC: Benjamin Herrenschmidt <benh at kernel.crashing.org>
> CC: Paul Mackerras <paulus at samba.org>
> CC: Christoph Lameter <cl at linux.com>
> Cc: Minchan Kim <minchan.kim at gmail.com>
> CC: Andrew Morton <akpm at linux-foundation.org>
> CC: KOSAKI Motohiro <kosaki.motohiro at jp.fujitsu.com>
> CC: Wen Congyang <wency at cn.fujitsu.com>
> Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki at jp.fujitsu.com>
>
> ---
>  drivers/base/memory.c  |   20 ++++++++++++++++++++
>  include/linux/memory.h |    1 +
>  mm/memory_hotplug.c    |    5 +++++
>  3 files changed, 26 insertions(+)
>
> Index: linux-3.5-rc4/drivers/base/memory.c
> ===================================================================
> --- linux-3.5-rc4.orig/drivers/base/memory.c    2012-06-26 13:28:16.726211752 +0900
> +++ linux-3.5-rc4/drivers/base/memory.c 2012-06-26 13:34:22.423639904 +0900
> @@ -70,6 +70,26 @@ void unregister_memory_isolate_notifier(
>  }
>  EXPORT_SYMBOL(unregister_memory_isolate_notifier);
>
> +bool memory_is_offline(unsigned long start_pfn, unsigned long end_pfn)

I dislike this function name. 'memory' is too vague to me.


> +{
> +       struct memory_block *mem;
> +       struct mem_section *section;
> +       unsigned long pfn, section_nr;
> +
> +       for (pfn = start_pfn; pfn < end_pfn; pfn += PAGES_PER_SECTION) {
> +               section_nr = pfn_to_section_nr(pfn);
> +               section = __nr_to_section(section_nr);
> +               mem = find_memory_block(section);

This seems to have strong sparse dependency.
Hm, I wonder why memory-hotplug.c can enable when X86_64_ACPI_NUMA.

# eventually, we can have this option just 'select SPARSEMEM'
config MEMORY_HOTPLUG
	bool "Allow for memory hot-add"
	depends on SPARSEMEM || X86_64_ACPI_NUMA


> +               if (!mem)
> +                       continue;
> +               if (mem->state == MEM_OFFLINE)
> +                       continue;
> +               return false;
> +       }
> +
> +       return true;
> +}
> +
>  /*
>  * register_memory - Setup a sysfs device for a memory block
>  */
> Index: linux-3.5-rc4/include/linux/memory.h
> ===================================================================
> --- linux-3.5-rc4.orig/include/linux/memory.h   2012-06-25 04:53:04.000000000 +0900
> +++ linux-3.5-rc4/include/linux/memory.h        2012-06-26 13:34:22.424639891 +0900
> @@ -120,6 +120,7 @@ extern int memory_isolate_notify(unsigne
>  extern struct memory_block *find_memory_block_hinted(struct mem_section *,
>                                                        struct memory_block *);
>  extern struct memory_block *find_memory_block(struct mem_section *);
> +extern bool memory_is_offline(unsigned long start_pfn, unsigned long end_pfn);
>  #define CONFIG_MEM_BLOCK_SIZE  (PAGES_PER_SECTION<<PAGE_SHIFT)
>  enum mem_add_context { BOOT, HOTPLUG };
>  #endif /* CONFIG_MEMORY_HOTPLUG_SPARSE */
> Index: linux-3.5-rc4/mm/memory_hotplug.c
> ===================================================================
> --- linux-3.5-rc4.orig/mm/memory_hotplug.c      2012-06-26 13:28:16.743211538 +0900
> +++ linux-3.5-rc4/mm/memory_hotplug.c   2012-06-26 13:48:38.264940468 +0900
> @@ -887,6 +887,11 @@ static int __ref offline_pages(unsigned
>
>        lock_memory_hotplug();
>
> +       if (memory_is_offline(start_pfn, end_pfn)) {
> +               ret = 0;
> +               goto out;
> +       }
> +
>        zone = page_zone(pfn_to_page(start_pfn));
>        node = zone_to_nid(zone);
>        nr_pages = end_pfn - start_pfn;
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo at kvack.org.  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont at kvack.org"> email at kvack.org </a>