linux-kernel - [PATCH v2 2/2] test_maple_tree: Add more testing for mas_empty

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]

Date:   Tue,  7 Mar 2023 13:02:47 -0500
From:   "Liam R. Howlett" <Liam.Howlett@...cle.com>
To:     maple-tree@...ts.infradead.org, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        Stable@...r.kernel.org, zhangpeng.00@...edance.com, snild@...y.com
Cc:     "Liam R. Howlett" <Liam.Howlett@...cle.com>
Subject: [PATCH v2 2/2] test_maple_tree: Add more testing for mas_empty_area()

Test robust filling of an entire area of the tree, then test one beyond.
This is to test the walking back up the tree at the end of nodes and
error condition.  Test inspired by the reproducer code provided by Snild
Dolkow.

The last test in the function tests for the case of a corrupted maple
state caused by the incorrect limits set during mas_skip_node().  There
needs to be a gap in the second last child and last child, but the
search must rule out the second last child's gap.  This would avoid
correcting the maple state to the correct max limit and return an error.

Cc: Snild Dolkow <snild@...y.com>
Link: https://lore.kernel.org/linux-mm/cb8dc31a-fef2-1d09-f133-e9f7b9f9e77a@sony.com/
Cc: <Stable@...r.kernel.org>
Cc: Peng Zhang <zhangpeng.00@...edance.com>
Fixes: e15e06a83923 ("lib/test_maple_tree: add testing for maple tree")
Signed-off-by: Liam R. Howlett <Liam.Howlett@...cle.com>
---
 lib/test_maple_tree.c | 48 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)

diff --git a/lib/test_maple_tree.c b/lib/test_maple_tree.c
index 3d19b1f78d71..f1db333270e9 100644
--- a/lib/test_maple_tree.c
+++ b/lib/test_maple_tree.c
@@ -2670,6 +2670,49 @@ static noinline void check_empty_area_window(struct maple_tree *mt)
 	rcu_read_unlock();
 }
 
+static noinline void check_empty_area_fill(struct maple_tree *mt)
+{
+	const unsigned long max = 0x25D78000;
+	unsigned long size;
+	int loop, shift;
+	MA_STATE(mas, mt, 0, 0);
+
+	mt_set_non_kernel(99999);
+	for (shift = 12; shift <= 16; shift++) {
+		loop = 5000;
+		size = 1 << shift;
+		while (loop--) {
+			mas_set(&mas, 0);
+			mas_lock(&mas);
+			MT_BUG_ON(mt, mas_empty_area(&mas, 0, max, size) != 0);
+			MT_BUG_ON(mt, mas.last != mas.index + size - 1);
+			mas_store_gfp(&mas, (void *)size, GFP_KERNEL);
+			mas_unlock(&mas);
+			mas_reset(&mas);
+		}
+	}
+
+	/* No space left. */
+	size = 0x1000;
+	rcu_read_lock();
+	MT_BUG_ON(mt, mas_empty_area(&mas, 0, max, size) != -EBUSY);
+	rcu_read_unlock();
+
+	/* Fill a depth 3 node to the maximum */
+	for (unsigned long i = 629440511; i <= 629440800; i += 6)
+		mtree_store_range(mt, i, i + 5, (void *)i, GFP_KERNEL);
+	/* Make space in the second-last depth 4 node */
+	mtree_erase(mt, 631668735);
+	/* Make space in the last depth 4 node */
+	mtree_erase(mt, 629506047);
+	mas_reset(&mas);
+	/* Search from just after the gap in the second-last depth 4 */
+	rcu_read_lock();
+	MT_BUG_ON(mt, mas_empty_area(&mas, 629506048, 690000000, 0x5000) != 0);
+	rcu_read_unlock();
+	mt_set_non_kernel(0);
+}
+
 static DEFINE_MTREE(tree);
 static int maple_tree_seed(void)
 {
@@ -2926,6 +2969,11 @@ static int maple_tree_seed(void)
 	check_empty_area_window(&tree);
 	mtree_destroy(&tree);
 
+	mt_init_flags(&tree, MT_FLAGS_ALLOC_RANGE);
+	check_empty_area_fill(&tree);
+	mtree_destroy(&tree);
+
+
 #if defined(BENCH)
 skip:
 #endif
-- 
2.39.2