linux-kernel - [PATCH] x86/mm: Fix {split,collapse}_page_count to use PTRS_PER

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 for Android: free password hash cracker in your pocket

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

Message-ID: <3d8bca84657d8b22a007a052258be1fb2a7c917d.1763531142.git.Yohei.Kojima@sony.com>
Date: Wed, 19 Nov 2025 14:46:19 +0900
From: Yohei Kojima <Yohei.Kojima@...y.com>
To: Dave Hansen <dave.hansen@...ux.intel.com>,
	Andy Lutomirski <luto@...nel.org>,
	Peter Zijlstra <peterz@...radead.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...hat.com>,
	Borislav Petkov <bp@...en8.de>,
	x86@...nel.org,
	"H. Peter Anvin" <hpa@...or.com>
Cc: Yohei Kojima <Yohei.Kojima@...y.com>,
	linux-kernel@...r.kernel.org
Subject: [PATCH] x86/mm: Fix {split,collapse}_page_count to use PTRS_PER_PMD if necessary

Before this commit, split_page_count() and collapse_page_count() updated
direct_pages_count using PTRS_PER_PTE constant. However, these functions
should use PTRS_PER_PMD if 1G page is splitted into 2M pages and vice
versa because 2M direct pages are managed by PMD.

This commit fixes {split,collapse}_page_count() to use PTRS_PER_PMD in
such cases. The basic behavior of these functions are unchanged because
x86's 1G page split and collapse are currently only supported on 64-bit
environment where PTRS_PER_PTE and PTRS_PER_PMD are both 512.
---
 arch/x86/mm/pat/set_memory.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index 970981893c9b..aa6fa4894edb 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -97,25 +97,29 @@ static void split_page_count(int level)
 		return;
 
 	direct_pages_count[level]--;
-	if (system_state == SYSTEM_RUNNING) {
-		if (level == PG_LEVEL_2M)
+	if (level == PG_LEVEL_2M) {
+		if (system_state == SYSTEM_RUNNING)
 			count_vm_event(DIRECT_MAP_LEVEL2_SPLIT);
-		else if (level == PG_LEVEL_1G)
+		direct_pages_count[PG_LEVEL_4K] += PTRS_PER_PTE;
+	} else if (level == PG_LEVEL_1G) {
+		if (system_state == SYSTEM_RUNNING)
 			count_vm_event(DIRECT_MAP_LEVEL3_SPLIT);
+		direct_pages_count[PG_LEVEL_2M] += PTRS_PER_PMD;
 	}
-	direct_pages_count[level - 1] += PTRS_PER_PTE;
 }
 
 static void collapse_page_count(int level)
 {
 	direct_pages_count[level]++;
-	if (system_state == SYSTEM_RUNNING) {
-		if (level == PG_LEVEL_2M)
+	if (level == PG_LEVEL_2M) {
+		if (system_state == SYSTEM_RUNNING)
 			count_vm_event(DIRECT_MAP_LEVEL2_COLLAPSE);
-		else if (level == PG_LEVEL_1G)
+		direct_pages_count[PG_LEVEL_4K] -= PTRS_PER_PTE;
+	} else if (level == PG_LEVEL_1G) {
+		if (system_state == SYSTEM_RUNNING)
 			count_vm_event(DIRECT_MAP_LEVEL3_COLLAPSE);
+		direct_pages_count[PG_LEVEL_2M] -= PTRS_PER_PMD;
 	}
-	direct_pages_count[level - 1] -= PTRS_PER_PTE;
 }
 
 void arch_report_meminfo(struct seq_file *m)
-- 
2.43.0