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.
Add a testcase to the maple tree test suite for this issue.
Cc: Andrei Vagin <[email protected]>
Cc: [email protected]
Reported-by: Mike Rapoport <[email protected]>
Fixes: 54a611b60590 ("Maple Tree: add new data structure")
Signed-off-by: Liam R. Howlett <[email protected]>
---
lib/maple_tree.c | 4 +++-
lib/test_maple_tree.c | 23 +++++++++++++++++++++++
2 files changed, 26 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);
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
On Fri, 16 Dec 2022 18:53:15 +0000 Liam Howlett <[email protected]> wrote:
> 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.
>
> Add a testcase to the maple tree test suite for this issue.
Shall do.
> Cc: Andrei Vagin <[email protected]>
> Cc: [email protected]
> Reported-by: Mike Rapoport <[email protected]>
> Fixes: 54a611b60590 ("Maple Tree: add new data structure")
> Signed-off-by: Liam R. Howlett <[email protected]>
I guess we want a cc:stable there?
> lib/test_maple_tree.c | 23 +++++++++++++++++++++++
Belated review: all this code runs at __init time, so every dang
function in there really should be marked __init, data marked
__initdata, etc. Like lib/test_bitmap.c.
I wonder if there's some trick we can do external to the .c file to
have the same effect.
Also, maple_tree_seed():set[] could be static ;). So we don't have to
initialize it at runtime. Better would be static const. Nitpick.
On Fri, Dec 16, 2022 at 06:53:15PM +0000, Liam Howlett wrote:
> 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.
>
> Add a testcase to the maple tree test suite for this issue.
>
> Cc: Andrei Vagin <[email protected]>
> Cc: [email protected]
> Reported-by: Mike Rapoport <[email protected]>
> Fixes: 54a611b60590 ("Maple Tree: add new data structure")
> Signed-off-by: Liam R. Howlett <[email protected]>
Tested-by: Mike Rapoport <[email protected]>
> ---
> lib/maple_tree.c | 4 +++-
> lib/test_maple_tree.c | 23 +++++++++++++++++++++++
> 2 files changed, 26 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);
>
> 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
--
Sincerely yours,
Mike.
* Andrew Morton <[email protected]> [221216 14:01]:
> On Fri, 16 Dec 2022 18:53:15 +0000 Liam Howlett <[email protected]> wrote:
>
> > 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.
> >
> > Add a testcase to the maple tree test suite for this issue.
>
> Shall do.
>
> > Cc: Andrei Vagin <[email protected]>
> > Cc: [email protected]
> > Reported-by: Mike Rapoport <[email protected]>
> > Fixes: 54a611b60590 ("Maple Tree: add new data structure")
> > Signed-off-by: Liam R. Howlett <[email protected]>
>
> I guess we want a cc:stable there?
Yes, please. I'll be sure to do that in the future for backports.
>
> > lib/test_maple_tree.c | 23 +++++++++++++++++++++++
>
> Belated review: all this code runs at __init time, so every dang
> function in there really should be marked __init, data marked
> __initdata, etc. Like lib/test_bitmap.c.
>
> I wonder if there's some trick we can do external to the .c file to
> have the same effect.
This can be built as a module and also runs in userspace via
tools/testing/radix-tree. I'm not sure about using __init in this
context but I'll look into it.
>
> Also, maple_tree_seed():set[] could be static ;). So we don't have to
> initialize it at runtime. Better would be static const. Nitpick.
I'll have a look. I'm sure there are more areas within my test code
that could be optimised like this.