Re: [PATCH 3/4] memory-hotplug: clear hwpoisoned flag when onliningpages

From: Ni zhan Chen
Date: Thu Sep 27 2012 - 08:28:05 EST

On 09/27/2012 01:45 PM, wency@xxxxxxxxxxxxxx wrote:
From: Wen Congyang <wency@xxxxxxxxxxxxxx>

hwpoisoned may set when we offline a page by the sysfs interface
/sys/devices/system/memory/soft_offline_page or
/sys/devices/system/memory/hard_offline_page. If we don't clear
this flag when onlining pages, this page can't be freed, and will
not in free list. So we can't offline these pages again. So we
should clear this flag when onlining pages.

page hwpoisoned maybe cause by a multi-bit ECC memory or cache failure, so this page should not be used, why you online and free it again? can any users use it?


CC: David Rientjes <rientjes@xxxxxxxxxx>
CC: Jiang Liu <liuj97@xxxxxxxxx>
CC: Len Brown <len.brown@xxxxxxxxx>
CC: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
CC: Paul Mackerras <paulus@xxxxxxxxx>
Cc: Minchan Kim <minchan.kim@xxxxxxxxx>
CC: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
CC: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
CC: Yasuaki Ishimatsu <isimatu.yasuaki@xxxxxxxxxxxxxx>
Signed-off-by: Wen Congyang <wency@xxxxxxxxxxxxxx>
---
mm/memory_hotplug.c | 8 ++++++++
1 files changed, 8 insertions(+), 0 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 6a5b90d..9a5b10f 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -431,6 +431,14 @@ EXPORT_SYMBOL_GPL(__online_page_increment_counters);
void __online_page_free(struct page *page)
{
+#ifdef CONFIG_MEMORY_FAILURE
+ /* The page may be marked HWPoisoned by soft/hard offline page */
+ if (PageHWPoison(page)) {
+ atomic_long_sub(1, &mce_bad_pages);
+ ClearPageHWPoison(page);
+ }
+#endif
+
ClearPageReserved(page);
init_page_count(page);
__free_page(page);

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/