[PATCH v2 0/2] maple_tree: Fix mas_spanning_rebalance() on

[PATCH v2 0/2] maple_tree: Fix mas_spanning_rebalance() on

[Liam Howlett]

Mike Rapoport contacted me off-list with a regression in running criu.
Periodic tests fail with an RCU stall during execution.  Although rare,
it is possible to hit this with other uses so this patch should be
backported to fix the regression.

This patch set adds the fix and a test case to the maple tree test
suite.

Link: https://lore.kernel.org/linux-mm/20221216185233.2036415-1-Liam.Howlett@oracle.com/

V1 changes:
- Add stable to CC list
- Add tested-by to Mike's Reported-by.
- Moved test case to its own commit since the test code was added after
  the fixed-by commit and isn't part of the fix.

Liam R. Howlett (2):
  maple_tree: Fix mas_spanning_rebalance() on insufficient data
  test_maple_tree: Add test for mas_spanning_rebalance() on insufficient
    data

 lib/maple_tree.c      |  4 +++-
 lib/test_maple_tree.c | 23 +++++++++++++++++++++++
 2 files changed, 26 insertions(+), 1 deletion(-)

-- 
2.35.1

[PATCH v2 1/2] maple_tree: Fix mas_spanning_rebalance() on insufficient data

[Liam Howlett]

Mike Rapoport contacted me off-list with a regression in running criu.
Periodic tests fail with an RCU stall during execution.  Although rare,
it is possible to hit this with other uses so this patch should be
backported to fix the regression.

An insufficient node was causing an out-of-bounds access on the node in
mas_leaf_max_gap().  The cause was the faulty detection of the new node
being a root node when overwriting many entries at the end of the tree.

Fix the detection of a new root and ensure there is sufficient data
prior to entering the spanning rebalance loop.

Cc: Andrei Vagin <avagin@gmail.com>
Cc: usama.anjum@collabora.com
CC: stable@vger.kernel.org
Reported-and-tested-by: Mike Rapoport <rppt@kernel.org>
Fixes: 54a611b60590 ("Maple Tree: add new data structure")
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
---
 lib/maple_tree.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lib/maple_tree.c b/lib/maple_tree.c
index fe3947b80069..26e2045d3cda 100644
--- a/lib/maple_tree.c
+++ b/lib/maple_tree.c
@@ -2994,7 +2994,9 @@ static int mas_spanning_rebalance(struct ma_state *mas,
 	mast->free = &free;
 	mast->destroy = &destroy;
 	l_mas.node = r_mas.node = m_mas.node = MAS_NONE;
-	if (!(mast->orig_l->min && mast->orig_r->max == ULONG_MAX) &&
+
+	/* Check if this is not root and has sufficient data.  */
+	if (((mast->orig_l->min != 0) || (mast->orig_r->max != ULONG_MAX)) &&
 	    unlikely(mast->bn->b_end <= mt_min_slots[mast->bn->type]))
 		mast_spanning_rebalance(mast);
 
-- 
2.35.1

[PATCH v2 2/2] test_maple_tree: Add test for mas_spanning_rebalance() on insufficient data

[Liam Howlett]

Add a test to the maple tree test suite for the spanning rebalance
insufficient node issue does not go undetected again.

Cc: Andrei Vagin <avagin@gmail.com>
Cc: usama.anjum@collabora.com
CC: stable@vger.kernel.org
Cc: Mike Rapoport <rppt@kernel.org>
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
---
 lib/test_maple_tree.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/lib/test_maple_tree.c b/lib/test_maple_tree.c
index f425f169ef08..497fc93ccf9e 100644
--- a/lib/test_maple_tree.c
+++ b/lib/test_maple_tree.c
@@ -2498,6 +2498,25 @@ static noinline void check_dup(struct maple_tree *mt)
 	}
 }
 
+static noinline void check_bnode_min_spanning(struct maple_tree *mt)
+{
+	int i = 50;
+	MA_STATE(mas, mt, 0, 0);
+
+	mt_set_non_kernel(9999);
+	mas_lock(&mas);
+	do {
+		mas_set_range(&mas, i*10, i*10+9);
+		mas_store(&mas, check_bnode_min_spanning);
+	} while (i--);
+
+	mas_set_range(&mas, 240, 509);
+	mas_store(&mas, NULL);
+	mas_unlock(&mas);
+	mas_destroy(&mas);
+	mt_set_non_kernel(0);
+}
+
 static DEFINE_MTREE(tree);
 static int maple_tree_seed(void)
 {
@@ -2742,6 +2761,10 @@ static int maple_tree_seed(void)
 	check_dup(&tree);
 	mtree_destroy(&tree);
 
+	mt_init_flags(&tree, MT_FLAGS_ALLOC_RANGE);
+	check_bnode_min_spanning(&tree);
+	mtree_destroy(&tree);
+
 #if defined(BENCH)
 skip:
 #endif
-- 
2.35.1