linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Wanpeng Li <liwanp@linux.vnet.ibm.com>
To: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	Andi Kleen <andi@firstfloor.org>,
	Fengguang Wu <fengguang.wu@intel.com>,
	Tony Luck <tony.luck@intel.com>,
	gong.chen@linux.intel.com, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH 4/6] mm/hwpoison: don't set migration type twice to avoid hold heavy contend zone->lock
Date: Fri, 23 Aug 2013 08:15:19 +0800	[thread overview]
Message-ID: <7264.85379641435$1377216937@news.gmane.org> (raw)
In-Reply-To: <1377198365-3xic0o2q-mutt-n-horiguchi@ah.jp.nec.com>

Hi Naoya,
On Thu, Aug 22, 2013 at 03:06:05PM -0400, Naoya Horiguchi wrote:
>On Thu, Aug 22, 2013 at 05:48:25PM +0800, Wanpeng Li wrote:
>> Set pageblock migration type will hold zone->lock which is heavy contended 
>> in system to avoid race. However, soft offline page will set pageblock 
>> migration type twice during get page if the page is in used, not hugetlbfs 
>> page and not on lru list. There is unnecessary to set the pageblock migration
>> type and hold heavy contended zone->lock again if the first round get page 
>> have already set the pageblock to right migration type.
>
>Can we use get_pageblock_migratetype() outside zone->lock?
>

I think the trick here is migration type is MIGRATE_ISOLATE. There 
are two parts can change MIGRATE_ISOLATE except hwpoison. One is 
memory hoplug, however, we hold lock_memory_hotplug() which avoid 
race. The second is CMA which umovable page allocation requst can't 
fallback to. So I think it's safe here.

Regards,
Wanpeng Li 

>There are surely some users which call this function outside
>zone->lock like free_hot_cold_pages(), __free_pages, etc.,
>but I think that there's a race window where migratetype is
>updated just after get_pageblock_migratetype() check.
>
>> Signed-off-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
>> ---
>>  mm/memory-failure.c | 3 ++-
>>  1 file changed, 2 insertions(+), 1 deletion(-)
>> 
>> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
>> index 6bfd51e..3bfb45f 100644
>> --- a/mm/memory-failure.c
>> +++ b/mm/memory-failure.c
>> @@ -1413,7 +1413,8 @@ static int __get_any_page(struct page *p, unsigned long pfn, int flags)
>>  	 * was free. This flag should be kept set until the source page
>>  	 * is freed and PG_hwpoison on it is set.
>>  	 */
>> -	set_migratetype_isolate(p, true);
>> +	if (get_pageblock_migratetype(p) == MIGRATE_ISOLATE)
>
>You meant '!=', right?
>
>Thanks,
>Naoya Horiguchi
>
>> +		set_migratetype_isolate(p, true);
>>  	/*
>>  	 * When the target page is a free hugepage, just remove it
>>  	 * from free hugepage list.
>> -- 
>> 1.8.1.2
>>
>
>--
>To unsubscribe, send a message with 'unsubscribe linux-mm' in
>the body to majordomo@kvack.org.  For more info on Linux MM,
>see: http://www.linux-mm.org/ .
>Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  parent reply	other threads:[~2013-08-23  0:15 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-08-22  9:48 [PATCH 1/6] mm/hwpoison: fix lose PG_dirty flag for errors on mlocked pages Wanpeng Li
2013-08-22  9:48 ` [PATCH 2/6] mm/hwpoison: don't need to hold compound lock for hugetlbfs page Wanpeng Li
2013-08-22 15:52   ` Naoya Horiguchi
2013-08-22 23:54     ` Wanpeng Li
2013-08-22 23:54     ` Wanpeng Li
2013-08-22  9:48 ` [PATCH 3/6] mm/hwpoison: fix num_poisoned_pages error statistics for thp Wanpeng Li
2013-08-22 16:43   ` Naoya Horiguchi
2013-08-22 17:00     ` Naoya Horiguchi
2013-08-22 23:52     ` Wanpeng Li
2013-08-22 23:52     ` Wanpeng Li
     [not found]     ` <5216a46f.a800310a.2351.ffffa95cSMTPIN_ADDED_BROKEN@mx.google.com>
2013-08-23  3:27       ` Naoya Horiguchi
2013-08-23  4:24         ` Wanpeng Li
2013-08-23  4:24         ` Wanpeng Li
2013-08-22  9:48 ` [PATCH 4/6] mm/hwpoison: don't set migration type twice to avoid hold heavy contend zone->lock Wanpeng Li
2013-08-22 19:06   ` Naoya Horiguchi
2013-08-23  0:15     ` Wanpeng Li
2013-08-23  0:15     ` Wanpeng Li [this message]
2013-08-22  9:48 ` [PATCH 5/6] mm/hwpoison: drop forward reference declarations __soft_offline_page() Wanpeng Li
2013-08-22 19:24   ` Naoya Horiguchi
2013-08-22  9:48 ` [PATCH 6/6] mm/hwpoison: centralize set PG_hwpoison flag and increase num_poisoned_pages Wanpeng Li
2013-08-22 20:13   ` Naoya Horiguchi
2013-08-23  0:03     ` Wanpeng Li
2013-08-23  0:03     ` Wanpeng Li
2013-08-22 15:51 ` [PATCH 1/6] mm/hwpoison: fix lose PG_dirty flag for errors on mlocked pages Naoya Horiguchi
2013-08-22 23:34   ` Wanpeng Li
2013-08-22 23:34   ` Wanpeng Li

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='7264.85379641435$1377216937@news.gmane.org' \
    --to=liwanp@linux.vnet.ibm.com \
    --cc=akpm@linux-foundation.org \
    --cc=andi@firstfloor.org \
    --cc=fengguang.wu@intel.com \
    --cc=gong.chen@linux.intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=n-horiguchi@ah.jp.nec.com \
    --cc=tony.luck@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox