From: Joonsoo Kim <iamjoonsoo.kim@lge.com>
To: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Mel Gorman <mgorman@suse.de>, Rik van Riel <riel@redhat.com>,
Rafael Aquini <aquini@redhat.com>,
Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>,
Christoph Lameter <cl@linux.com>,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
Vlastimil Babka <vbabka@suse.cz>,
Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Subject: Re: [PATCH v2 2/7] mm/migrate: correct failure handling if !hugepage_migration_support()
Date: Tue, 10 Dec 2013 17:45:07 +0900 [thread overview]
Message-ID: <20131210084507.GC24992@lge.com> (raw)
In-Reply-To: <52a679dd.8886420a.34d4.1b0eSMTPIN_ADDED_BROKEN@mx.google.com>
On Tue, Dec 10, 2013 at 10:17:56AM +0800, Wanpeng Li wrote:
> Hi Joonsoo,
> On Mon, Dec 09, 2013 at 06:10:43PM +0900, Joonsoo Kim wrote:
> >We should remove the page from the list if we fail without ENOSYS,
> >since migrate_pages() consider error cases except -ENOMEM and -EAGAIN
> >as permanent failure and it assumes that the page would be removed from
> >the list. Without this patch, we could overcount number of failure.
> >
> >In addition, we should put back the new hugepage if
> >!hugepage_migration_support(). If not, we would leak hugepage memory.
> >
> >Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
> >
> >diff --git a/mm/migrate.c b/mm/migrate.c
> >index c6ac87a..b1cfd01 100644
> >--- a/mm/migrate.c
> >+++ b/mm/migrate.c
> >@@ -1011,7 +1011,7 @@ static int unmap_and_move_huge_page(new_page_t get_new_page,
> > {
> > int rc = 0;
> > int *result = NULL;
> >- struct page *new_hpage = get_new_page(hpage, private, &result);
> >+ struct page *new_hpage;
> > struct anon_vma *anon_vma = NULL;
> >
> > /*
> >@@ -1021,9 +1021,12 @@ static int unmap_and_move_huge_page(new_page_t get_new_page,
> > * tables or check whether the hugepage is pmd-based or not before
> > * kicking migration.
> > */
> >- if (!hugepage_migration_support(page_hstate(hpage)))
> >+ if (!hugepage_migration_support(page_hstate(hpage))) {
> >+ putback_active_hugepage(hpage);
> > return -ENOSYS;
>
> The memory hotplug(do_migrate_range) and hwpoison(soft_offline_huge_page) callsets both
> will call putback_movable_pages/putback_active_hugepage for -ENOSYS case.
Hello Wanpeng.
Yes, those callsite handle error case, but error case handling should be done
in unmap_and_move_huge_page(). It is mentioned on patch 1. If we defer to
remove the pages from the list, nr_failed can be overcounted.
Thanks.
--
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>
next prev parent reply other threads:[~2013-12-10 8:42 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-12-09 9:10 [PATCH v2 0/7] correct and clean-up migration related stuff Joonsoo Kim
2013-12-09 9:10 ` [PATCH v2 1/7] mm/migrate: add comment about permanent failure path Joonsoo Kim
2013-12-10 1:42 ` Wanpeng Li
2013-12-09 9:10 ` [PATCH v2 2/7] mm/migrate: correct failure handling if !hugepage_migration_support() Joonsoo Kim
2013-12-09 16:17 ` Christoph Lameter
2013-12-10 8:38 ` Joonsoo Kim
2013-12-10 2:17 ` Wanpeng Li
2013-12-10 8:45 ` Joonsoo Kim [this message]
2013-12-10 9:27 ` Wanpeng Li
2013-12-09 9:10 ` [PATCH v2 3/7] mm/mempolicy: correct putback method for isolate pages if failed Joonsoo Kim
2013-12-09 16:41 ` Rafael Aquini
2013-12-10 2:22 ` Wanpeng Li
2013-12-09 9:10 ` [PATCH v2 4/7] mm/migrate: remove putback_lru_pages, fix comment on putback_movable_pages Joonsoo Kim
2013-12-10 2:33 ` Wanpeng Li
2013-12-10 8:51 ` Joonsoo Kim
2013-12-10 9:00 ` Joonsoo Kim
2013-12-09 9:10 ` [PATCH v2 5/7] mm/compaction: respect ignore_skip_hint in update_pageblock_skip Joonsoo Kim
2013-12-10 2:50 ` Wanpeng Li
2013-12-09 9:10 ` [PATCH v2 6/7] mm/migrate: remove unused function, fail_migrate_page() Joonsoo Kim
2013-12-09 16:20 ` Christoph Lameter
2013-12-09 17:07 ` Naoya Horiguchi
2013-12-10 2:36 ` Wanpeng Li
2013-12-09 9:10 ` [PATCH v2 7/7] mm/migrate: remove result argument on page allocation function for migration Joonsoo Kim
2013-12-09 16:40 ` Christoph Lameter
2013-12-10 3:22 ` Wanpeng Li
2013-12-11 8:47 ` Joonsoo Kim
2013-12-11 16:00 ` Christoph Lameter
2013-12-12 0:07 ` Joonsoo Kim
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=20131210084507.GC24992@lge.com \
--to=iamjoonsoo.kim@lge.com \
--cc=akpm@linux-foundation.org \
--cc=aquini@redhat.com \
--cc=cl@linux.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=liwanp@linux.vnet.ibm.com \
--cc=mgorman@suse.de \
--cc=n-horiguchi@ah.jp.nec.com \
--cc=riel@redhat.com \
--cc=vbabka@suse.cz \
--cc=zhangyanfei@cn.fujitsu.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;
as well as URLs for NNTP newsgroup(s).