From: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
To: Matthew Wilcox <willy@infradead.org>
Cc: Michal Hocko <mhocko@kernel.org>,
"Kirill A. Shutemov" <kirill@shutemov.name>,
Zi Yan <zi.yan@sent.com>,
"linux-mm@kvack.org" <linux-mm@kvack.org>,
Andrew Morton <akpm@linux-foundation.org>,
Vlastimil Babka <vbabka@suse.cz>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] mm: shmem: enable thp migration (Re: [PATCH v1] mm: consider non-anonymous thp as unmovable page)
Date: Mon, 23 Apr 2018 07:21:02 +0000 [thread overview]
Message-ID: <20180423072101.GA12157@hori1.linux.bs1.fc.nec.co.jp> (raw)
In-Reply-To: <20180423030349.GB2308@bombadil.infradead.org>
On Sun, Apr 22, 2018 at 08:03:49PM -0700, Matthew Wilcox wrote:
> On Fri, Apr 06, 2018 at 03:07:11AM +0000, Naoya Horiguchi wrote:
> > Subject: [PATCH] mm: enable thp migration for shmem thp
>
> This patch is buggy, but not in a significant way:
>
> > @@ -524,13 +524,26 @@ int migrate_page_move_mapping(struct address_space *mapping,
> > }
> >
> > radix_tree_replace_slot(&mapping->i_pages, pslot, newpage);
>
> ^^^ this line should have been deleted
>
> > + if (PageTransHuge(page)) {
> > + int i;
> > + int index = page_index(page);
> > +
> > + for (i = 0; i < HPAGE_PMD_NR; i++) {
> ^^^ or this iteration should start at 1
> > + pslot = radix_tree_lookup_slot(&mapping->i_pages,
> > + index + i);
> > + radix_tree_replace_slot(&mapping->i_pages, pslot,
> > + newpage + i);
> > + }
> > + } else {
> > + radix_tree_replace_slot(&mapping->i_pages, pslot, newpage);
> ^^^ and if the second option, then we don't need this line
> > + }
>
> So either this:
>
> - radix_tree_replace_slot(&mapping->i_pages, pslot, newpage);
> + if (PageTransHuge(page)) {
> + int i;
> + int index = page_index(page);
> +
> + for (i = 0; i < HPAGE_PMD_NR; i++) {
> + pslot = radix_tree_lookup_slot(&mapping->i_pages,
> + index + i);
> + radix_tree_replace_slot(&mapping->i_pages, pslot,
> + newpage + i);
> + }
> + } else {
> + radix_tree_replace_slot(&mapping->i_pages, pslot, newpage);
> + }
>
> Or this:
>
> radix_tree_replace_slot(&mapping->i_pages, pslot, newpage);
> + if (PageTransHuge(page)) {
> + int i;
> + int index = page_index(page);
> +
> + for (i = 1; i < HPAGE_PMD_NR; i++) {
> + pslot = radix_tree_lookup_slot(&mapping->i_pages,
> + index + i);
> + radix_tree_replace_slot(&mapping->i_pages, pslot,
> + newpage + i);
> + }
> + }
>
> The second one is shorter and involves fewer lookups ...
Hi Matthew,
Thank you for poinitng out, I like the second one.
The original patch is now in upsteam, so I wrote a patch on it.
Thanks,
Naoya
--------
From: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Date: Sun, 22 Apr 2018 20:03:49 -0700
Subject: [PATCH] mm: migrate: fix double call of radix_tree_replace_slot()
radix_tree_replace_slot() is called twice for head page, it's
obviously a bug. Let's fix it.
Fixes: e71769ae5260 ("mm: enable thp migration for shmem thp")
Reported-by: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
---
mm/migrate.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/mm/migrate.c b/mm/migrate.c
index 568433023831..8c0af0f7cab1 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -528,14 +528,12 @@ int migrate_page_move_mapping(struct address_space *mapping,
int i;
int index = page_index(page);
- for (i = 0; i < HPAGE_PMD_NR; i++) {
+ for (i = 1; i < HPAGE_PMD_NR; i++) {
pslot = radix_tree_lookup_slot(&mapping->i_pages,
index + i);
radix_tree_replace_slot(&mapping->i_pages, pslot,
newpage + i);
}
- } else {
- radix_tree_replace_slot(&mapping->i_pages, pslot, newpage);
}
/*
--
2.7.4
next prev parent reply other threads:[~2018-04-23 7:23 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-04-03 4:46 [PATCH v1] mm: consider non-anonymous thp as unmovable page Naoya Horiguchi
2018-04-03 7:59 ` Michal Hocko
2018-04-03 8:24 ` Naoya Horiguchi
2018-04-03 8:34 ` Michal Hocko
2018-04-03 10:54 ` Kirill A. Shutemov
2018-04-03 10:58 ` Michal Hocko
2018-04-03 11:16 ` Kirill A. Shutemov
2018-04-03 11:33 ` Michal Hocko
2018-04-05 8:59 ` Michal Hocko
2018-04-05 12:28 ` Kirill A. Shutemov
2018-04-05 12:48 ` Michal Hocko
2018-04-05 13:40 ` Kirill A. Shutemov
2018-04-05 15:05 ` Michal Hocko
2018-04-05 15:55 ` Kirill A. Shutemov
2018-04-05 16:03 ` Michal Hocko
2018-04-05 17:58 ` Zi Yan
2018-04-05 19:04 ` Michal Hocko
2018-04-05 19:09 ` Zi Yan
2018-04-06 3:07 ` [PATCH] mm: shmem: enable thp migration (Re: [PATCH v1] mm: consider non-anonymous thp as unmovable page) Naoya Horiguchi
2018-04-06 5:14 ` Naoya Horiguchi
2018-04-06 7:08 ` Michal Hocko
2018-04-09 7:14 ` Naoya Horiguchi
2018-04-10 11:16 ` Kirill A. Shutemov
2018-04-11 9:26 ` Michal Hocko
2018-04-11 19:27 ` Andrew Morton
2018-04-11 19:43 ` Michal Hocko
2018-04-23 3:03 ` Matthew Wilcox
2018-04-23 7:21 ` Naoya Horiguchi [this message]
2018-04-05 16:42 ` [PATCH v1] mm: consider non-anonymous thp as unmovable page Sasha Levin
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=20180423072101.GA12157@hori1.linux.bs1.fc.nec.co.jp \
--to=n-horiguchi@ah.jp.nec.com \
--cc=akpm@linux-foundation.org \
--cc=kirill@shutemov.name \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@kernel.org \
--cc=vbabka@suse.cz \
--cc=willy@infradead.org \
--cc=zi.yan@sent.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