Re: [PATCH bpf-next v3 2/2] selftests/bpf: Add a negative test for stack accounting in jit mode

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




On 2/13/24 11:40 AM, Andrii Nakryiko wrote:
On Thu, Feb 8, 2024 at 1:54 PM Yonghong Song <yonghong.song@xxxxxxxxx> wrote:
The new test is very similar to test_global_func1.c, but
is modified to fail on jit mode.

Signed-off-by: Yonghong Song <yonghong.song@xxxxxxxxx>
---
  .../bpf/prog_tests/test_global_funcs.c        |  3 ++
  .../selftests/bpf/progs/test_global_func18.c  | 44 +++++++++++++++++++
  2 files changed, 47 insertions(+)
  create mode 100644 tools/testing/selftests/bpf/progs/test_global_func18.c

diff --git a/tools/testing/selftests/bpf/prog_tests/test_global_funcs.c b/tools/testing/selftests/bpf/prog_tests/test_global_funcs.c
index a3a41680b38e..dccbf2213135 100644
--- a/tools/testing/selftests/bpf/prog_tests/test_global_funcs.c
+++ b/tools/testing/selftests/bpf/prog_tests/test_global_funcs.c
@@ -18,6 +18,7 @@
  #include "test_global_func15.skel.h"
  #include "test_global_func16.skel.h"
  #include "test_global_func17.skel.h"
+#include "test_global_func18.skel.h"
  #include "test_global_func_ctx_args.skel.h"

  #include "bpf/libbpf_internal.h"
@@ -140,6 +141,8 @@ void test_test_global_funcs(void)
  {
         if (!env.jit_enabled) {
                 RUN_TESTS(test_global_func1);
+       } else {
+               RUN_TESTS(test_global_func18);
         }

         RUN_TESTS(test_global_func2);
diff --git a/tools/testing/selftests/bpf/progs/test_global_func18.c b/tools/testing/selftests/bpf/progs/test_global_func18.c
new file mode 100644
index 000000000000..d1aa3b2c68fe
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/test_global_func18.c
@@ -0,0 +1,44 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (c) 2024 Meta Platforms, Inc. and affiliates. */
+#include <stddef.h>
+#include <linux/bpf.h>
+#include <bpf/bpf_helpers.h>
+#include "bpf_misc.h"
+
+#define MAX_STACK1 (512 - 3 * 32 + 8)
+#define MAX_STACK2 (3 * 32)
+
+__attribute__ ((noinline))
nit: we have __noinline defined, let's use it as a less verbose option?

A copy-paste issue. In v4, I will drop this patch
since I will increase the stack size to accommodate
both jit and !jit cases.


+int f1(struct __sk_buff *skb)
+{
+       return skb->len;
+}
+
+int f3(int, struct __sk_buff *skb, int);
+
+__attribute__ ((noinline))
+int f2(int val, struct __sk_buff *skb)
+{
+       volatile char buf[MAX_STACK1] = {};
+
+       __sink(buf[MAX_STACK1 - 1]);
+
+       return f1(skb) + f3(val, skb, 1);
+}
+
+__attribute__ ((noinline))
+int f3(int val, struct __sk_buff *skb, int var)
+{
+       volatile char buf[MAX_STACK2] = {};
+
+       __sink(buf[MAX_STACK2 - 1]);
+
+       return skb->ifindex * val * var;
+}
+
+SEC("tc")
+__failure __msg("combined stack size of 3 calls is 528")
+int global_func18(struct __sk_buff *skb)
+{
+       return f1(skb) + f2(2, skb) + f3(3, skb, 4);
+}
--
2.39.3






[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux