linux-kernel - Re: [PATCH bpf-next] test_bpf: add module parameter test

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>] [day] [month] [year] [list]
Message-ID: <CAM1=_QTb8zejZTqnAaobhTErkowBm=p-fuiLYwDFtWYoUuNGXQ@mail.gmail.com>
Date:   Fri, 1 Oct 2021 15:51:14 +0200
From:   Johan Almbladh <johan.almbladh@...finetworks.com>
To:     Tiezhu Yang <yangtiezhu@...ngson.cn>
Cc:     Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andrii@...nel.org>,
        Networking <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>,
        linux-kernel@...r.kernel.org, Xuefeng Li <lixuefeng@...ngson.cn>
Subject: Re: [PATCH bpf-next] test_bpf: add module parameter test_type

Hi Tiezhu,

Your v2 is base64-encoded. Please use plain-text for patch submissions.

On Thu, Sep 30, 2021 at 1:01 PM Tiezhu Yang <yangtiezhu@...ngson.cn> wrote:
>
> After commit 9298e63eafea ("bpf/tests: Add exhaustive tests of ALU
> operand magnitudes"), when modprobe test_bpf.ko with jit on mips64,
> there exists segment fault due to fhe following reason:
>
> test_bpf: #616 ALU64_MOV_X: all register value magnitudes jited:1
> Break instruction in kernel code[#1]
>
> It seems that the related jit implementations of some test cases
> in test_bpf() have problems. At this moment, I do not care about
> the segment fault while I just want to verify the test cases of
> tail calls.

Don't put too much effort into the current MIPS64 JIT. I have been
working on a significant upgrade of the MIPS JIT, which adds MIPS32
support and full eBPF ISA support, among other things. All the new JIT
tests in test_bpf.ko I submitted are essentially a side effect of that
work.

I am currently testing the new JIT on different setups, and I hope to
be able to submit the patch set next week. A side note, as you seem to
work at Loongson. It would be great if you could verify the CPU errata
workarounds I implemented for Loongson-2F and 3, once I get the patch
set out for review.

>
> Based on the above background and motivation, add the following
> module parameter test_type to the test_bpf.ko:
> test_type=<string>: only the specified type will be run, the string
> can be "test_bpf", "test_tail_calls" or "test_skb_segment".
>
> This is useful to only test the corresponding test type when specify
> the valid test_type string.

I agree that it is good to be able to choose a particular test suite
to run. There are also the test_id and test_range parameters. If we
add a test suite selector, it would be nice if the test range/id
selection applied to that test suite, instead of being ignored for all
suites except test_bpf.

>
> Any invalid test type will result in -EINVAL being returned and no
> tests being run. If the test_type is not specified or specified as
> empty string, it does not change the current logic, all of the test
> cases will be run.
>
> Signed-off-by: Tiezhu Yang <yangtiezhu@...ngson.cn>
> ---
>  lib/test_bpf.c | 48 +++++++++++++++++++++++++++++++++---------------
>  1 file changed, 33 insertions(+), 15 deletions(-)
>
> diff --git a/lib/test_bpf.c b/lib/test_bpf.c
> index 21ea1ab..9428fec 100644
> --- a/lib/test_bpf.c
> +++ b/lib/test_bpf.c
> @@ -11866,6 +11866,9 @@ module_param(test_id, int, 0);
>  static int test_range[2] = { 0, ARRAY_SIZE(tests) - 1 };
>  module_param_array(test_range, int, NULL, 0);
>
> +static char test_type[32];
> +module_param_string(test_type, test_type, sizeof(test_type), 0);
> +
>  static __init int find_test_index(const char *test_name)
>  {
>         int i;
> @@ -12518,24 +12521,39 @@ static int __init test_bpf_init(void)
>         struct bpf_array *progs = NULL;
>         int ret;
>
> -       ret = prepare_bpf_tests();
> -       if (ret < 0)
> -               return ret;
> +       if (strlen(test_type) &&
> +           strcmp(test_type, "test_bpf") &&
> +           strcmp(test_type, "test_tail_calls") &&
> +           strcmp(test_type, "test_skb_segment")) {
> +               pr_err("test_bpf: invalid test_type '%s' specified.\n", test_type);
> +               return -EINVAL;
> +       }
> +
> +       if (!strlen(test_type) || !strcmp(test_type, "test_bpf")) {
> +               ret = prepare_bpf_tests();
> +               if (ret < 0)
> +                       return ret;
> +
> +               ret = test_bpf();
> +               destroy_bpf_tests();
> +               if (ret)
> +                       return ret;
> +       }
>
> -       ret = test_bpf();
> -       destroy_bpf_tests();
> -       if (ret)
> -               return ret;
> +       if (!strlen(test_type) || !strcmp(test_type, "test_tail_calls")) {
> +               ret = prepare_tail_call_tests(&progs);
> +               if (ret)
> +                       return ret;
> +               ret = test_tail_calls(progs);
> +               destroy_tail_call_tests(progs);
> +               if (ret)
> +                       return ret;
> +       }
>
> -       ret = prepare_tail_call_tests(&progs);
> -       if (ret)
> -               return ret;
> -       ret = test_tail_calls(progs);
> -       destroy_tail_call_tests(progs);
> -       if (ret)
> -               return ret;
> +       if (!strlen(test_type) || !strcmp(test_type, "test_skb_segment"))
> +               return test_skb_segment();
>
> -       return test_skb_segment();
> +       return 0;
>  }
>
>  static void __exit test_bpf_exit(void)
> --
> 2.1.0
>