[PATCH v3 1/2] bpf: Tighten cgroup storage cookie checks for prog arrays

[PATCH v3 1/2] bpf: Tighten cgroup storage cookie checks for prog arrays

[Lin Ma]

The recent KCTF-reported cgroup local storage issue assigned
CVE-2025-38502 was fixed by commit abad3d0bad72 ("bpf: Fix oob access
in cgroup local storage").

However, the previous fixes are still incomplete. The current prog-array
compatibility check treats a program with no cgroup storage as
compatible with any stored storage cookie. This allows a storage-less
program to bridge a tail-call chain between an entry program and a
storage-using callee even though runtime cgroup local storage still
follows the caller context.

Require exact per-type storage_cookie equality when checking prog-array
compatibility. This blocks zero-storage bridge programs from joining a
prog-array owned by a storage-using program and closes the residual
A -> B(no storage) -> C(storage) path.

This also aligns with Amery Hung's earlier NULL-storage tail-call fix by
requiring storage use to match consistently across prog-array users.

Cc: stable@vger.kernel.org
Fixes: abad3d0bad72 ("bpf: Fix oob access in cgroup local storage")
Tested-by: Amery Hung <ameryhung@gmail.com>
Signed-off-by: Lin Ma <malin89@huawei.com>
Signed-off-by: Rongzhen Cui <cuirongzhen@huawei.com>
Signed-off-by: Jingguo Tan <tanjingguo@huawei.com>
---
v1: https://lore.kernel.org/bpf/20260601095158.1186318-1-malin89@huawei.com/
v1 -> v2:
 - refine the commit message and mention the relation to Amery Hung's
   NULL-storage tail-call fix
 - add patch 2/2 selftests for tail-call cgroup storage prog-array
   checks
v2: https://lore.kernel.org/bpf/31927f33-9db0-4a39-b38a-72b33487979e@linux.dev/T/#t
v2 -> v3:
 - use abad3d0bad72 as the Fixes tag
---
 kernel/bpf/core.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index 6aa2a8b24030..f0b61b10f30e 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -2470,8 +2470,12 @@ static bool __bpf_prog_map_compatible(struct bpf_map *map,
 				break;
 			cookie = aux->cgroup_storage[i] ?
 				 aux->cgroup_storage[i]->cookie : 0;
-			ret = map->owner->storage_cookie[i] == cookie ||
-			      !cookie;
+			/*
+			 * Tail calls keep using the caller cgroup storage
+			 * context, so prog-array members must use the same
+			 * storage cookie.
+			 */
+			ret = map->owner->storage_cookie[i] == cookie;
 		}
 		if (ret &&
 		    map->owner->attach_func_proto != aux->attach_func_proto) {
-- 
2.53.0

[PATCH v3 2/2] selftests/bpf: Cover tail-call cgroup storage prog-array checks

[Lin Ma]

Add tail-call selftests for prog-array ownership when cgroup storage is
in use. Verify that loading succeeds when callers and callees reuse the
owner's cgroup storage map, and that loading fails for a different
storage map and for the A(storage) -> B(no storage) -> C(storage)
bridge case from patch 1.

Signed-off-by: Lin Ma <malin89@huawei.com>
Signed-off-by: Rongzhen Cui <cuirongzhen@huawei.com>
Signed-off-by: Jingguo Tan <tanjingguo@huawei.com>
---
v2: https://lore.kernel.org/bpf/31927f33-9db0-4a39-b38a-72b33487979e@linux.dev/T/#t
v2 -> v3:
 - add explicit A(storage) -> B(no storage) -> C(storage) bridge
   coverage
---
 .../selftests/bpf/prog_tests/tailcalls.c      | 159 ++++++++++++++++++
 .../bpf/progs/tailcall_cgrp_storage.c         |  44 +++++
 .../progs/tailcall_cgrp_storage_no_storage.c  |  20 +++
 .../bpf/progs/tailcall_cgrp_storage_owner.c   |  32 ++++
 4 files changed, 255 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/progs/tailcall_cgrp_storage.c
 create mode 100644 tools/testing/selftests/bpf/progs/tailcall_cgrp_storage_no_storage.c
 create mode 100644 tools/testing/selftests/bpf/progs/tailcall_cgrp_storage_owner.c

diff --git a/tools/testing/selftests/bpf/prog_tests/tailcalls.c b/tools/testing/selftests/bpf/prog_tests/tailcalls.c
index 7d534fde0af9..f8c4b569da95 100644
--- a/tools/testing/selftests/bpf/prog_tests/tailcalls.c
+++ b/tools/testing/selftests/bpf/prog_tests/tailcalls.c
@@ -8,6 +8,9 @@
 #include "tailcall_freplace.skel.h"
 #include "tc_bpf2bpf.skel.h"
 #include "tailcall_fail.skel.h"
+#include "tailcall_cgrp_storage_owner.skel.h"
+#include "tailcall_cgrp_storage_no_storage.skel.h"
+#include "tailcall_cgrp_storage.skel.h"
 #include "tailcall_sleepable.skel.h"
 
 /* test_tailcall_1 checks basic functionality by patching multiple locations
@@ -1654,6 +1657,154 @@ static void test_tailcall_failure()
 	RUN_TESTS(tailcall_fail);
 }
 
+static void test_tailcall_cgrp_storage(void)
+{
+	struct tailcall_cgrp_storage_owner *owner_skel = NULL;
+	struct tailcall_cgrp_storage *skel = NULL;
+	int err, key = 0, prog_array_fd, prog_fd, storage_map_fd;
+
+	owner_skel = tailcall_cgrp_storage_owner__open_and_load();
+	if (!ASSERT_OK_PTR(owner_skel, "owner_open_and_load"))
+		return;
+
+	prog_array_fd = bpf_map__fd(owner_skel->maps.prog_array);
+	storage_map_fd = bpf_map__fd(owner_skel->maps.storage_map);
+
+	skel = tailcall_cgrp_storage__open();
+	if (!ASSERT_OK_PTR(skel, "tailcall_cgrp_storage__open"))
+		goto out;
+
+	err = bpf_map__reuse_fd(skel->maps.prog_array, prog_array_fd);
+	if (!ASSERT_OK(err, "reuse_prog_array"))
+		goto out;
+
+	err = bpf_map__reuse_fd(skel->maps.storage_map, storage_map_fd);
+	if (!ASSERT_OK(err, "reuse_storage_map"))
+		goto out;
+
+	err = bpf_object__load(skel->obj);
+	if (!ASSERT_OK(err, "tailcall_cgrp_storage__load"))
+		goto out;
+
+	prog_fd = bpf_program__fd(skel->progs.callee_prog);
+	err = bpf_map_update_elem(prog_array_fd, &key, &prog_fd, BPF_ANY);
+	ASSERT_OK(err, "update_prog_array");
+
+out:
+	tailcall_cgrp_storage__destroy(skel);
+	tailcall_cgrp_storage_owner__destroy(owner_skel);
+}
+
+static void test_tailcall_cgrp_storage_diff_storage(void)
+{
+	struct tailcall_cgrp_storage_owner *owner_skel = NULL;
+	struct tailcall_cgrp_storage *skel = NULL;
+	int err, prog_array_fd;
+
+	owner_skel = tailcall_cgrp_storage_owner__open_and_load();
+	if (!ASSERT_OK_PTR(owner_skel, "owner_open_and_load"))
+		return;
+
+	prog_array_fd = bpf_map__fd(owner_skel->maps.prog_array);
+
+	skel = tailcall_cgrp_storage__open();
+	if (!ASSERT_OK_PTR(skel, "tailcall_cgrp_storage__open"))
+		goto out;
+
+	err = bpf_map__reuse_fd(skel->maps.prog_array, prog_array_fd);
+	if (!ASSERT_OK(err, "reuse_prog_array"))
+		goto out;
+
+	err = bpf_object__load(skel->obj);
+	ASSERT_ERR(err, "tailcall_cgrp_storage__load");
+
+out:
+	tailcall_cgrp_storage__destroy(skel);
+	tailcall_cgrp_storage_owner__destroy(owner_skel);
+}
+
+static void test_tailcall_cgrp_storage_no_storage(void)
+{
+	struct tailcall_cgrp_storage_owner *owner_skel = NULL;
+	struct tailcall_cgrp_storage_no_storage *skel = NULL;
+	int err, prog_array_fd;
+
+	owner_skel = tailcall_cgrp_storage_owner__open_and_load();
+	if (!ASSERT_OK_PTR(owner_skel, "owner_open_and_load"))
+		return;
+
+	prog_array_fd = bpf_map__fd(owner_skel->maps.prog_array);
+
+	skel = tailcall_cgrp_storage_no_storage__open();
+	if (!ASSERT_OK_PTR(skel, "tailcall_cgrp_storage_no_storage__open"))
+		goto out;
+
+	err = bpf_map__reuse_fd(skel->maps.prog_array, prog_array_fd);
+	if (!ASSERT_OK(err, "reuse_prog_array"))
+		goto out;
+
+	err = bpf_object__load(skel->obj);
+	ASSERT_ERR(err, "tailcall_cgrp_storage_no_storage__load");
+
+out:
+	tailcall_cgrp_storage_no_storage__destroy(skel);
+	tailcall_cgrp_storage_owner__destroy(owner_skel);
+}
+
+static void test_tailcall_cgrp_storage_no_storage_bridge(void)
+{
+	struct tailcall_cgrp_storage_owner *owner_skel = NULL;
+	struct tailcall_cgrp_storage_no_storage *bridge_skel = NULL;
+	struct tailcall_cgrp_storage *callee_skel = NULL;
+	int err, key = 0, prog_array_fd, prog_fd, storage_map_fd;
+
+	owner_skel = tailcall_cgrp_storage_owner__open_and_load();
+	if (!ASSERT_OK_PTR(owner_skel, "owner_open_and_load"))
+		return;
+
+	prog_array_fd = bpf_map__fd(owner_skel->maps.prog_array);
+	storage_map_fd = bpf_map__fd(owner_skel->maps.storage_map);
+
+	callee_skel = tailcall_cgrp_storage__open();
+	if (!ASSERT_OK_PTR(callee_skel, "tailcall_cgrp_storage__open"))
+		goto out;
+
+	bpf_program__set_autoload(callee_skel->progs.caller_prog, false);
+
+	err = bpf_map__reuse_fd(callee_skel->maps.prog_array, prog_array_fd);
+	if (!ASSERT_OK(err, "reuse_prog_array"))
+		goto out;
+
+	err = bpf_map__reuse_fd(callee_skel->maps.storage_map, storage_map_fd);
+	if (!ASSERT_OK(err, "reuse_storage_map"))
+		goto out;
+
+	err = bpf_object__load(callee_skel->obj);
+	if (!ASSERT_OK(err, "tailcall_cgrp_storage__load"))
+		goto out;
+
+	prog_fd = bpf_program__fd(callee_skel->progs.callee_prog);
+	err = bpf_map_update_elem(prog_array_fd, &key, &prog_fd, BPF_ANY);
+	if (!ASSERT_OK(err, "update_prog_array"))
+		goto out;
+
+	bridge_skel = tailcall_cgrp_storage_no_storage__open();
+	if (!ASSERT_OK_PTR(bridge_skel, "tailcall_cgrp_storage_no_storage__open"))
+		goto out;
+
+	err = bpf_map__reuse_fd(bridge_skel->maps.prog_array, prog_array_fd);
+	if (!ASSERT_OK(err, "reuse_prog_array"))
+		goto out;
+
+	err = bpf_object__load(bridge_skel->obj);
+	ASSERT_ERR(err, "tailcall_cgrp_storage_no_storage_bridge__load");
+
+out:
+	tailcall_cgrp_storage_no_storage__destroy(bridge_skel);
+	tailcall_cgrp_storage__destroy(callee_skel);
+	tailcall_cgrp_storage_owner__destroy(owner_skel);
+}
+
 noinline void uprobe_sleepable_trigger(void)
 {
 	asm volatile ("");
@@ -1777,6 +1928,14 @@ void test_tailcalls(void)
 		test_tailcall_freplace();
 	if (test__start_subtest("tailcall_bpf2bpf_freplace"))
 		test_tailcall_bpf2bpf_freplace();
+	if (test__start_subtest("tailcall_cgrp_storage"))
+		test_tailcall_cgrp_storage();
+	if (test__start_subtest("tailcall_cgrp_storage_diff_storage"))
+		test_tailcall_cgrp_storage_diff_storage();
+	if (test__start_subtest("tailcall_cgrp_storage_no_storage"))
+		test_tailcall_cgrp_storage_no_storage();
+	if (test__start_subtest("tailcall_cgrp_storage_no_storage_bridge"))
+		test_tailcall_cgrp_storage_no_storage_bridge();
 	if (test__start_subtest("tailcall_failure"))
 		test_tailcall_failure();
 	if (test__start_subtest("tailcall_sleepable"))
diff --git a/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage.c b/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage.c
new file mode 100644
index 000000000000..4dd3a0033d75
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage.c
@@ -0,0 +1,44 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE);
+	__type(key, struct bpf_cgroup_storage_key);
+	__type(value, __u64);
+} storage_map SEC(".maps");
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PROG_ARRAY);
+	__uint(max_entries, 1);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(__u32));
+} prog_array SEC(".maps");
+
+SEC("cgroup_skb/egress")
+int caller_prog(struct __sk_buff *skb)
+{
+	__u64 *storage;
+
+	storage = bpf_get_local_storage(&storage_map, 0);
+	if (storage)
+		*storage = 1;
+
+	bpf_tail_call(skb, &prog_array, 0);
+	return 1;
+}
+
+SEC("cgroup_skb/egress")
+int callee_prog(struct __sk_buff *skb)
+{
+	__u64 *storage;
+
+	storage = bpf_get_local_storage(&storage_map, 0);
+	if (storage)
+		*storage = 1;
+
+	return 1;
+}
+
+char _license[] SEC("license") = "GPL";
diff --git a/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage_no_storage.c b/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage_no_storage.c
new file mode 100644
index 000000000000..f95166dad7c5
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage_no_storage.c
@@ -0,0 +1,20 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PROG_ARRAY);
+	__uint(max_entries, 1);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(__u32));
+} prog_array SEC(".maps");
+
+SEC("cgroup_skb/egress")
+int caller_prog(struct __sk_buff *skb)
+{
+	bpf_tail_call(skb, &prog_array, 0);
+	return 1;
+}
+
+char _license[] SEC("license") = "GPL";
diff --git a/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage_owner.c b/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage_owner.c
new file mode 100644
index 000000000000..d7e8ec9855c5
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/tailcall_cgrp_storage_owner.c
@@ -0,0 +1,32 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE);
+	__type(key, struct bpf_cgroup_storage_key);
+	__type(value, __u64);
+} storage_map SEC(".maps");
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PROG_ARRAY);
+	__uint(max_entries, 1);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(__u32));
+} prog_array SEC(".maps");
+
+SEC("cgroup_skb/egress")
+int prog_array_owner(struct __sk_buff *skb)
+{
+	__u64 *storage;
+
+	storage = bpf_get_local_storage(&storage_map, 0);
+	if (storage)
+		*storage = 1;
+
+	bpf_tail_call(skb, &prog_array, 0);
+	return 1;
+}
+
+char _license[] SEC("license") = "GPL";
-- 
2.53.0

Re: [PATCH v3 1/2] bpf: Tighten cgroup storage cookie checks for prog arrays

[bot+bpf-ci]

[-- Attachment #1: Type: text/plain, Size: 1510 bytes --]

> diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> index a656a8572bdb..d85869544dc8 100644
> --- a/kernel/bpf/core.c
> +++ b/kernel/bpf/core.c
> @@ -2480,8 +2480,12 @@ static bool __bpf_prog_map_compatible(struct bpf_map *map,
>  				break;
>  			cookie = aux->cgroup_storage[i] ?
>  				 aux->cgroup_storage[i]->cookie : 0;
> -			ret = map->owner->storage_cookie[i] == cookie ||
> -			      !cookie;
> +			/*
> +			 * Tail calls keep using the caller cgroup storage
> +			 * context, so prog-array members must use the same
> +			 * storage cookie.
> +			 */
> +			ret = map->owner->storage_cookie[i] == cookie;
>  		}
>  		if (ret &&
>  		    map->owner->attach_func_proto != aux->attach_func_proto) {

Related to this patch series, Leon Hwang raised a concern about test
placement in the selftest file (tools/testing/selftests/bpf/prog_tests/
tailcalls.c). The new cgroup storage tail-call tests should be placed at
the tail of test_tailcalls() rather than before tailcall_failure and
tailcall_sleepable.

Reference: https://lore.kernel.org/bpf/20260603150213.68746-1-leon.hwang@linux.dev/

Leon's comment from v2: "These new ones should stay at the tail of
test_tailcalls()."

Should the test ordering be adjusted in the selftest file to address this
feedback?


---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/26923843925

Re: [PATCH v3 2/2] selftests/bpf: Cover tail-call cgroup storage prog-array checks

[bot+bpf-ci]

[-- Attachment #1: Type: text/plain, Size: 1717 bytes --]

> diff --git a/tools/testing/selftests/bpf/prog_tests/tailcalls.c b/tools/testing/selftests/bpf/prog_tests/tailcalls.c
> index 7d534fde0af9..f8c4b569da95 100644
> --- a/tools/testing/selftests/bpf/prog_tests/tailcalls.c
> +++ b/tools/testing/selftests/bpf/prog_tests/tailcalls.c

[ ... ]

> @@ -1777,6 +1928,14 @@ void test_tailcalls(void)
>  		test_tailcall_freplace();
>  	if (test__start_subtest("tailcall_bpf2bpf_freplace"))
>  		test_tailcall_bpf2bpf_freplace();
> +	if (test__start_subtest("tailcall_cgrp_storage"))
> +		test_tailcall_cgrp_storage();
> +	if (test__start_subtest("tailcall_cgrp_storage_diff_storage"))
> +		test_tailcall_cgrp_storage_diff_storage();
> +	if (test__start_subtest("tailcall_cgrp_storage_no_storage"))
> +		test_tailcall_cgrp_storage_no_storage();
> +	if (test__start_subtest("tailcall_cgrp_storage_no_storage_bridge"))
> +		test_tailcall_cgrp_storage_no_storage_bridge();
>  	if (test__start_subtest("tailcall_failure"))
>  		test_tailcall_failure();
>  	if (test__start_subtest("tailcall_sleepable"))
>  		test_tailcall_sleepable();

In the v2 review Leon Hwang asked that these new subtests be placed at the
tail of test_tailcalls():

  https://lore.kernel.org/bpf/20260603150213.68746-1-leon.hwang@linux.dev/

  "These new ones should stay at the tail of test_tailcalls()."

Here they are still registered ahead of tailcall_failure and
tailcall_sleepable rather than at the end.  Was the intent to move them
after the existing subtests?


---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/26923843925