From: Miaohe Lin <linmiaohe@huawei.com>
To: "Huang, Ying" <ying.huang@intel.com>
Cc: <akpm@linux-foundation.org>, <linux-mm@kvack.org>,
<linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 2/8] mm/vmscan: remove unneeded can_split_huge_page check
Date: Thu, 31 Mar 2022 18:47:49 +0800 [thread overview]
Message-ID: <0e84a88d-b97e-6e3b-2d26-07b957d8bcdd@huawei.com> (raw)
In-Reply-To: <877d8abjoj.fsf@yhuang6-desk2.ccr.corp.intel.com>
On 2022/3/31 16:38, Huang, Ying wrote:
> Miaohe Lin <linmiaohe@huawei.com> writes:
>
>> can_split_huge_page is introduced via commit b8f593cd0896 ("mm, THP, swap:
>> check whether THP can be split firstly") to avoid deleting the THP from
>> the swap cache and freeing the swap cluster when the THP cannot be split.
>> But since commit bd4c82c22c36 ("mm, THP, swap: delay splitting THP after
>> swapped out"), splitting THP is delayed until THP is swapped out. There's
>> no need to delete the THP from the swap cache and free the swap cluster
>> anymore. Thus we can remove this unneeded can_split_huge_page check now to
>> simplify the code logic.
>>
>> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
>> ---
>> mm/vmscan.c | 3 ---
>> 1 file changed, 3 deletions(-)
>>
>> diff --git a/mm/vmscan.c b/mm/vmscan.c
>> index 7c1a9713bfc9..09b452c4d256 100644
>> --- a/mm/vmscan.c
>> +++ b/mm/vmscan.c
>> @@ -1691,9 +1691,6 @@ static unsigned int shrink_page_list(struct list_head *page_list,
>> if (folio_maybe_dma_pinned(folio))
>> goto keep_locked;
>> if (PageTransHuge(page)) {
>> - /* cannot split THP, skip it */
>> - if (!can_split_folio(folio, NULL))
>> - goto activate_locked;
>> /*
>> * Split pages without a PMD map right
>> * away. Chances are some or all of the
>
> I'm OK with the change itself. But THP still needs to be split after
> being swapped out. The reason we don't need to check can_split_folio()
Could you please explain the relevant code path more detailed slightly?
IIUC, if THP is swapped out, it will be freed via destroy_compound_page
after __remove_mapping in shrink_page_list. So THP can be freed without
split. Or am I miss something ?
> is that folio_maybe_dma_pinned() is checked before. Which will avoid
> the long term pinned pages to be swapped out. And we can live with
> short term pinned pages. Without can_split_folio() checking we can
> simplify the code as follows,
>
> /*
> * Split pages without a PMD map right
> * away. Chances are some or all of the
> * tail pages can be freed without IO.
> */
> if (PageTransHuge(page) && !compound_mapcount(page) &&
> split_huge_page_to_list(page, page_list))
> goto keep_locked;
>
> activate_locked can be changed to keep_locked too, because it's just
> short term pinning.
>
The change above looks good to me. Many thanks. Should I add a Suggested-by
tag for you?
> Best Regards,
> Huang, Ying
Thanks.
>
> .
>
next prev parent reply other threads:[~2022-03-31 10:47 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-29 13:26 [PATCH 0/8] A few cleanup and fixup patches for migration Miaohe Lin
2022-03-29 13:26 ` [PATCH 1/8] mm/vmscan: remove redundant folio_test_swapbacked check when folio is file lru Miaohe Lin
2022-03-29 13:46 ` Matthew Wilcox
2022-03-30 1:46 ` Miaohe Lin
2022-03-30 8:13 ` Muchun Song
2022-03-30 9:26 ` Miaohe Lin
2022-03-31 6:37 ` Huang, Ying
2022-03-31 7:44 ` Miaohe Lin
2022-03-31 8:02 ` Huang, Ying
2022-03-31 8:07 ` Miaohe Lin
2022-03-29 13:26 ` [PATCH 2/8] mm/vmscan: remove unneeded can_split_huge_page check Miaohe Lin
2022-03-31 8:38 ` Huang, Ying
2022-03-31 10:47 ` Miaohe Lin [this message]
2022-03-29 13:26 ` [PATCH 3/8] mm/vmscan: introduce helper function reclaim_page_list() Miaohe Lin
2022-03-29 13:50 ` Matthew Wilcox
2022-03-30 2:04 ` Miaohe Lin
2022-03-30 16:23 ` Matthew Wilcox
2022-03-31 1:45 ` Miaohe Lin
2022-03-29 14:00 ` Christoph Hellwig
2022-03-30 2:02 ` Miaohe Lin
2022-03-29 13:26 ` [PATCH 4/8] mm/vmscan: save a bit of stack space in shrink_lruvec Miaohe Lin
2022-03-29 14:00 ` Christoph Hellwig
2022-03-30 1:59 ` Miaohe Lin
2022-03-29 13:26 ` [PATCH 5/8] mm/vmscan: use helper folio_is_file_lru() Miaohe Lin
2022-04-01 3:20 ` Huang, Ying
2022-04-01 6:14 ` Miaohe Lin
2022-04-01 6:40 ` Huang, Ying
2022-03-29 13:26 ` [PATCH 6/8] mm/vmscan: take all base pages of THP into account when race with speculative reference Miaohe Lin
2022-03-29 13:26 ` [PATCH 7/8] mm/vmscan: take min_slab_pages into account when try to call shrink_node Miaohe Lin
2022-03-29 13:26 ` [PATCH 8/8] mm/vmscan: remove obsolete comment in kswapd_run Miaohe Lin
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=0e84a88d-b97e-6e3b-2d26-07b957d8bcdd@huawei.com \
--to=linmiaohe@huawei.com \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=ying.huang@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;
as well as URLs for NNTP newsgroup(s).