linux-kernel - [PATCH] mm/page_alloc: optimize lowmem_reserve max lookup using monotonicity

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <tencent_ED235B379E160A5C2BCF688ADDF3921EC808@qq.com>
Date: Fri, 14 Nov 2025 10:40:32 +0000
From: fujunjie <fujunjie1@...com>
To: akpm@...ux-foundation.org
Cc: vbabka@...e.cz,
	surenb@...gle.com,
	mhocko@...e.com,
	jackmanb@...gle.com,
	hannes@...xchg.org,
	ziy@...dia.com,
	linux-mm@...ck.org,
	linux-kernel@...r.kernel.org,
	fujunjie <fujunjie1@...com>
Subject: [PATCH] mm/page_alloc: optimize lowmem_reserve max lookup using monotonicity

calculate_totalreserve_pages() currently finds the maximum
lowmem_reserve[j] for a zone by scanning the full range
[j = zone_idx .. MAX_NR_ZONES). However,
setup_per_zone_lowmem_reserve() constructs lowmem_reserve[]
monotonically increasing in j for a fixed zone (and never populates
lowmem_reserve[zone_idx] itself). This means the maximum valid reserve
entry always resides at the highest j > zone_idx that has a non-zero value.

Rewrite the loop to walk backwards from MAX_NR_ZONES - 1 down to
zone_idx + 1, stopping at the first non-zero lowmem_reserve[j]. This
avoids a full-range scan and makes the intent clearer. Behavior remains
unchanged.

Although this code is not on a hot path, the revised form is clearer
and avoids an unnecessary full scan.

Signed-off-by: fujunjie <fujunjie1@...com>
---
 mm/page_alloc.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 600d9e981c23d..414c5ba978418 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6285,10 +6285,22 @@ static void calculate_totalreserve_pages(void)
 			long max = 0;
 			unsigned long managed_pages = zone_managed_pages(zone);

-			/* Find valid and maximum lowmem_reserve in the zone */
-			for (j = i; j < MAX_NR_ZONES; j++)
-				max = max(max, zone->lowmem_reserve[j]);
+			/*
+			 * Find valid and maximum lowmem_reserve in the zone.
+			 *
+			 * setup_per_zone_lowmem_reserve() builds
+			 * lowmem_reserve[j] monotonically increasing in j
+			 * for a fixed zone, so the maximum lives at the
+			 * highest index that has a non-zero value.  Walk
+			 * backwards and stop at the first hit.
+			 */
+			for (j = MAX_NR_ZONES - 1; j > i; j--) {
+				if (!zone->lowmem_reserve[j])
+					continue;

+				max = zone->lowmem_reserve[j];
+				break;
+			}
 			/* we treat the high watermark as reserved pages. */
 			max += high_wmark_pages(zone);

-- 
2.34.1