Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Zi Yan <zi.yan@sent.com>
To:     linux-mm@kvack.org, Matthew Wilcox <willy@infradead.org>
Cc:     "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>,
        Roman Gushchin <guro@fb.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
        Yang Shi <shy828301@gmail.com>,
        Michal Hocko <mhocko@kernel.org>,
        John Hubbard <jhubbard@nvidia.com>,
        Ralph Campbell <rcampbell@nvidia.com>,
        David Nellans <dnellans@nvidia.com>, Zi Yan <ziy@nvidia.com>
Subject: [RFC PATCH 5/6] mm: truncate: split thp to a non-zero order if possible.
Date:   Wed, 11 Nov 2020 15:40:07 -0500
Message-Id: <20201111204008.21332-6-zi.yan@sent.com>
In-Reply-To: <20201111204008.21332-1-zi.yan@sent.com>
References: <20201111204008.21332-1-zi.yan@sent.com>
Reply-To: Zi Yan <ziy@nvidia.com>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

From: Zi Yan <ziy@nvidia.com>

To minimize the number of pages after a truncation, when truncating a
THP, we do not need to split it all the way down to order-0. The THP has
at most three parts, the part before offset, the part to be truncated,
the part left at the end. Use the non-zero minimum of them to decide
what order we split the THP to.

Signed-off-by: Zi Yan <ziy@nvidia.com>
---
 mm/truncate.c | 22 ++++++++++++++++++++--
 1 file changed, 20 insertions(+), 2 deletions(-)

diff --git a/mm/truncate.c b/mm/truncate.c
index 20bd17538ec2..6d8e3c6115bc 100644
--- a/mm/truncate.c
+++ b/mm/truncate.c
@@ -237,7 +237,7 @@ int truncate_inode_page(struct address_space *mapping, =
struct page *page)
 bool truncate_inode_partial_page(struct page *page, loff_t start, loff_t e=
nd)
 {
 	loff_t pos =3D page_offset(page);
-	unsigned int offset, length;
+	unsigned int offset, length, left, min_subpage_size =3D PAGE_SIZE;
=20
 	if (pos < start)
 		offset =3D start - pos;
@@ -248,6 +248,7 @@ bool truncate_inode_partial_page(struct page *page, lof=
f_t start, loff_t end)
 		length =3D length - offset;
 	else
 		length =3D end + 1 - pos - offset;
+	left =3D thp_size(page) - offset - length;
=20
 	wait_on_page_writeback(page);
 	if (length =3D=3D thp_size(page)) {
@@ -267,7 +268,24 @@ bool truncate_inode_partial_page(struct page *page, lo=
ff_t start, loff_t end)
 		do_invalidatepage(page, offset, length);
 	if (!PageTransHuge(page))
 		return true;
-	return split_huge_page(page) =3D=3D 0;
+
+	/*
+	 * find the non-zero minimum of offset, length, and left and use it to
+	 * decide the new order of the page after split
+	 */
+	if (offset && left)
+		min_subpage_size =3D min_t(unsigned int,
+					 min_t(unsigned int, offset, length),
+					 left);
+	else if (!offset)
+		min_subpage_size =3D min_t(unsigned int, length, left);
+	else /* !left */
+		min_subpage_size =3D min_t(unsigned int, length, offset);
+
+	min_subpage_size =3D max_t(unsigned int, PAGE_SIZE, min_subpage_size);
+
+	return split_huge_page_to_list_to_order(page, NULL,
+				ilog2(min_subpage_size/PAGE_SIZE)) =3D=3D 0;
 }
=20
 /*
--=20
2.28.0