Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Fri, 10 Apr 2020 13:26:18 -0700
From:   Kees Cook <keescook@chromium.org>
To:     Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
Cc:     bpf@vger.kernel.org, netdev@vger.kernel.org,
        linux-kselftest@vger.kernel.org, linux-kernel@vger.kernel.org,
        luto@amacapital.net, wad@chromium.org, shuah@kernel.org
Subject: Re: [PATCH] selftests/seccomp: allow clock_nanosleep instead of
 nanosleep
Message-ID: <202004101325.CF69610F77@keescook>
References: <20200408235753.8566-1-cascardo@canonical.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200408235753.8566-1-cascardo@canonical.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Apr 08, 2020 at 08:57:53PM -0300, Thadeu Lima de Souza Cascardo wrote:
> glibc 2.31 calls clock_nanosleep when its nanosleep function is used. So
> the restart_syscall fails after that. In order to deal with it, we trace
> clock_nanosleep and nanosleep. Then we check for either.
> 
> This works just fine on systems with both glibc 2.30 and glibc 2.31,
> whereas it failed before on a system with glibc 2.31.
> 
> Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>

Thanks for this! I'm trying to determine if all architectures have
__NR_clock_nanosleep ... got some test builds running now, but if it all
builds fine, then I'll get this sent to Linus for -rc2.

-Kees

> ---
>  tools/testing/selftests/seccomp/seccomp_bpf.c | 14 ++++++++------
>  1 file changed, 8 insertions(+), 6 deletions(-)
> 
> diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c
> index 89fb3e0b552e..c0aa46ce14f6 100644
> --- a/tools/testing/selftests/seccomp/seccomp_bpf.c
> +++ b/tools/testing/selftests/seccomp/seccomp_bpf.c
> @@ -2803,12 +2803,13 @@ TEST(syscall_restart)
>  			 offsetof(struct seccomp_data, nr)),
>  
>  #ifdef __NR_sigreturn
> -		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_sigreturn, 6, 0),
> +		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_sigreturn, 7, 0),
>  #endif
> -		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_read, 5, 0),
> -		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_exit, 4, 0),
> -		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_rt_sigreturn, 3, 0),
> -		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_nanosleep, 4, 0),
> +		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_read, 6, 0),
> +		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_exit, 5, 0),
> +		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_rt_sigreturn, 4, 0),
> +		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_nanosleep, 5, 0),
> +		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_clock_nanosleep, 4, 0),
>  		BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_restart_syscall, 4, 0),
>  
>  		/* Allow __NR_write for easy logging. */
> @@ -2895,7 +2896,8 @@ TEST(syscall_restart)
>  	ASSERT_EQ(PTRACE_EVENT_SECCOMP, (status >> 16));
>  	ASSERT_EQ(0, ptrace(PTRACE_GETEVENTMSG, child_pid, NULL, &msg));
>  	ASSERT_EQ(0x100, msg);
> -	EXPECT_EQ(__NR_nanosleep, get_syscall(_metadata, child_pid));
> +	ret = get_syscall(_metadata, child_pid);
> +	EXPECT_TRUE(ret == __NR_nanosleep || ret == __NR_clock_nanosleep);
>  
>  	/* Might as well check siginfo for sanity while we're here. */
>  	ASSERT_EQ(0, ptrace(PTRACE_GETSIGINFO, child_pid, NULL, &info));
> -- 
> 2.20.1
> 

-- 
Kees Cook