linux-kernel - [PATCH RFC 2/9] mm/page_owner: add parameter option 'page

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

Message-ID: <20251205231721.104505-3-mfo@igalia.com>
Date: Fri,  5 Dec 2025 20:17:14 -0300
From: Mauricio Faria de Oliveira <mfo@...lia.com>
To: Andrew Morton <akpm@...ux-foundation.org>,
	David Hildenbrand <david@...nel.org>
Cc: Lorenzo Stoakes <lorenzo.stoakes@...cle.com>,
	Michal Hocko <mhocko@...e.com>,
	Vlastimil Babka <vbabka@...e.cz>,
	Oscar Salvador <osalvador@...e.de>,
	linux-mm@...ck.org,
	linux-kernel@...r.kernel.org,
	kernel-dev@...lia.com
Subject: [PATCH RFC 2/9] mm/page_owner: add parameter option 'page_owner=on,swap'

CONFIG_SWAP_PAGE_OWNER will introduce a behavior change in page_owner on
systems with swap space. Thus, add the 'swap' option to the 'page_owner'
parameter to enable it ('page_owner=on,swap').

Currently, the allocation stack trace of a page is lost at swap-out (the
page is freed) and reset at swap-in (a new page is allocated).

With the new feature, if this option is enabled, the (initial) allocation
stack trace of a page is maintained over swap-out and swap-in, and the
refcounts for the initial and swap-in allocation stack traces are adjusted.

(Note: 'swap_page_owner_enabled' is parsed before 'page_owner_enabled', so
'page_owner=off,swap' could set the former but not the latter, but this is
a nop as then init_page_owner() returns early and doesn't enable anything.)

Signed-off-by: Mauricio Faria de Oliveira <mfo@...lia.com>
---
 mm/page_owner.c | 22 +++++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

diff --git a/mm/page_owner.c b/mm/page_owner.c
index a70245684206..a27958872b07 100644
--- a/mm/page_owner.c
+++ b/mm/page_owner.c
@@ -57,6 +57,11 @@ struct stack_print_ctx {
 static bool page_owner_enabled __initdata;
 DEFINE_STATIC_KEY_FALSE(page_owner_inited);
 
+#ifdef CONFIG_SWAP_PAGE_OWNER
+static bool swap_page_owner_enabled __initdata;
+DEFINE_STATIC_KEY_FALSE(swap_page_owner_inited);
+#endif
+
 static depot_stack_handle_t dummy_handle;
 static depot_stack_handle_t failure_handle;
 static depot_stack_handle_t early_handle;
@@ -81,7 +86,18 @@ static inline void unset_current_in_page_owner(void)
 
 static int __init early_page_owner_param(char *buf)
 {
-	int ret = kstrtobool(buf, &page_owner_enabled);
+	int ret;
+
+#ifdef CONFIG_SWAP_PAGE_OWNER
+	char *swap = strchr(buf, ',');
+
+	if (swap) {
+		*swap++ = '\0';
+		if (!strcmp(swap, "swap"))
+			swap_page_owner_enabled = true;
+	}
+#endif
+	ret = kstrtobool(buf, &page_owner_enabled);
 
 	if (page_owner_enabled)
 		stack_depot_request_early_init();
@@ -138,6 +154,10 @@ static __init void init_page_owner(void)
 	dummy_stack.next = &failure_stack;
 	stack_list = &dummy_stack;
 	static_branch_enable(&page_owner_inited);
+#ifdef CONFIG_SWAP_PAGE_OWNER
+	if (swap_page_owner_enabled)
+		static_branch_enable(&swap_page_owner_inited);
+#endif
 }
 
 struct page_ext_operations page_owner_ops = {
-- 
2.51.0