linux-kernel - [PATCH v2 07/13] mm: thp: make split queue lock safe when LRU pages reparented

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: <20210916134748.67712-8-songmuchun@bytedance.com>
Date:   Thu, 16 Sep 2021 21:47:42 +0800
From:   Muchun Song <songmuchun@...edance.com>
To:     guro@...com, hannes@...xchg.org, mhocko@...nel.org,
        akpm@...ux-foundation.org, shakeelb@...gle.com,
        vdavydov.dev@...il.com
Cc:     linux-kernel@...r.kernel.org, linux-mm@...ck.org,
        duanxiongchun@...edance.com, fam.zheng@...edance.com,
        bsingharora@...il.com, shy828301@...il.com, alexs@...nel.org,
        smuchun@...il.com, zhengqi.arch@...edance.com,
        Muchun Song <songmuchun@...edance.com>
Subject: [PATCH v2 07/13] mm: thp: make split queue lock safe when LRU pages reparented

Similar to lruvec lock, we use the same approach to make the split
queue lock safe when LRU pages reparented.

Signed-off-by: Muchun Song <songmuchun@...edance.com>
---
 mm/huge_memory.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 9d8dfa82991a..12950d4988e6 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -535,9 +535,22 @@ static struct deferred_split *split_queue_lock(struct page *head)
 {
 	struct deferred_split *queue;
 
+	rcu_read_lock();
+retry:
 	queue = page_split_queue(head);
 	spin_lock(&queue->split_queue_lock);
 
+	if (unlikely(split_queue_memcg(queue) != page_memcg(head))) {
+		spin_unlock(&queue->split_queue_lock);
+		goto retry;
+	}
+
+	/*
+	 * Preemption is disabled in the internal of spin_lock, which can serve
+	 * as RCU read-side critical sections.
+	 */
+	rcu_read_unlock();
+
 	return queue;
 }
 
@@ -546,9 +559,19 @@ split_queue_lock_irqsave(struct page *head, unsigned long *flags)
 {
 	struct deferred_split *queue;
 
+	rcu_read_lock();
+retry:
 	queue = page_split_queue(head);
 	spin_lock_irqsave(&queue->split_queue_lock, *flags);
 
+	if (unlikely(split_queue_memcg(queue) != page_memcg(head))) {
+		spin_unlock_irqrestore(&queue->split_queue_lock, *flags);
+		goto retry;
+	}
+
+	/* See the comments in split_queue_lock(). */
+	rcu_read_unlock();
+
 	return queue;
 }
 
-- 
2.11.0