Re: Re: [PATCH] selftests/bpf: Fix strncpy() fortify warning

From: Andrii Nakryiko <andrii.nakryiko@gmail.com>
To: Rong Tao <rtoax@foxmail.com>
Cc: andrii@kernel.org, ast@kernel.org, bpf@vger.kernel.org,
	daniel@iogearbox.net, haoluo@google.com,
	john.fastabend@gmail.com, jolsa@kernel.org, kpsingh@kernel.org,
	linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
	martin.lau@linux.dev, mykolal@fb.com, rongtao@cestc.cn,
	sdf@google.com, shuah@kernel.org, song@kernel.org, yhs@fb.com
Subject: Re: Re: [PATCH] selftests/bpf: Fix strncpy() fortify warning
Date: Fri, 28 Oct 2022 10:08:25 -0700	[thread overview]
Message-ID: <CAEf4BzazYVkVKrKzPD8a7tRZrcWDvvgoVksJHYk3+46V=8kZhw@mail.gmail.com> (raw)
In-Reply-To: <tencent_CD22930180507496CE65B30ACBABC4681506@qq.com>

On Thu, Oct 27, 2022 at 5:26 PM Rong Tao <rtoax@foxmail.com> wrote:
>
> Thanks for your reply, `enable[0] = '\0';` at the beginning and then
> strncat() still has the same compile warning
>
> --- a/tools/testing/selftests/bpf/cgroup_helpers.c
> +++ b/tools/testing/selftests/bpf/cgroup_helpers.c
> @@ -77,7 +77,8 @@ static int __enable_controllers(const char *cgroup_path, const char *controllers
>                 enable[len] = 0;
>                 close(fd);
>         } else {
> -               strncpy(enable, controllers, sizeof(enable));
> +               enable[0] = '\0';
> +               strncat(enable, controllers, sizeof(enable));
>         }
>
> In function ‘__enable_controllers’:
> tools/testing/selftests/bpf/cgroup_helpers.c:81:17: warning: ‘strncat’ specified bound 4097 equals destination size [-Wstringop-truncation]
>    81 |                 strncat(enable, controllers, sizeof(enable));
>       |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> tools/testing/selftests/bpf/cgroup_helpers.c:81:17: warning: ‘strncat’ specified bound 4097 equals destination size [-Wstringop-overflow=]
>
> So, i think just add '-1' for strncpy() is a good way.

no, it's not, see my previous email about ending up with
non-zero-terminated C string.

check strncat() API, it leaves the dst string zero terminated, and
yes, you need -1 for strncat as well, your compiler is right