linux-kernel - Re: [PATCH] perf bench: fix assert when NDEBUG is defined

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>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <87txvfev6o.fsf@sejong.aot.lge.com>
Date:	Mon, 03 Sep 2012 10:45:03 +0900
From:	Namhyung Kim <namhyung@...nel.org>
To:	Irina Tirdea <irina.tirdea@...il.com>
Cc:	Steven Rostedt <rostedt@...dmis.org>,
	Arnaldo Carvalho de Melo <acme@...stprotocols.net>,
	Ingo Molnar <mingo@...nel.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Frederic Weisbecker <fweisbec@...il.com>
Subject: Re: [PATCH] perf bench: fix assert when NDEBUG is defined

On Mon, 3 Sep 2012 03:04:32 +0300, Irina Tirdea wrote:
> From: Irina Tirdea <irina.tirdea@...el.com>
>
> When NDEBUG is defined, the assert macro will be expanded to nothing.
> Some assert calls used in perf are also including some functionality
> (e.g. system calls), not only validity checks. Therefore, if NDEBUG is
> defined, these functionality will be removed along with the assert.
>
> The functionality of the program needs to be separated from the assert checks.
> In perf, BUG_ON is also defined on assert, so we need to fix these statements
> too.
>
> Signed-off-by: Irina Tirdea <irina.tirdea@...el.com>
> ---
>  tools/perf/bench/mem-memcpy.c |    8 +++++---
>  tools/perf/bench/mem-memset.c |    8 +++++---
>  tools/perf/bench/sched-pipe.c |    6 ++++--
>  3 files changed, 14 insertions(+), 8 deletions(-)
>
> diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c
> index 02dad5d..bccb783 100644
> --- a/tools/perf/bench/mem-memcpy.c
> +++ b/tools/perf/bench/mem-memcpy.c
> @@ -144,17 +144,19 @@ static double do_memcpy_gettimeofday(memcpy_t
> fn, size_t len, bool prefault)
>  {
>  	struct timeval tv_start, tv_end, tv_diff;
>  	void *src = NULL, *dst = NULL;
> -	int i;
> +	int i, ret;
>
>  	alloc_mem(&src, &dst, len);
>
>  	if (prefault)
>  		fn(dst, src, len);
>
> -	BUG_ON(gettimeofday(&tv_start, NULL));
> +	ret = gettimeofday(&tv_start, NULL);
> +	BUG_ON(ret);

I think one of good thing of assert is that it outputs the exact failure
condition when it fails.  So with patch, it will convert

  Assertion `gettimeofday(&tv_start, NULL)' failed.

into

  Assertion `ret' failed.

which is not so informative.

So I'd rather suggest using more descriptive names like ret_gtod ?


>  	for (i = 0; i < iterations; ++i)
>  		fn(dst, src, len);
> -	BUG_ON(gettimeofday(&tv_end, NULL));
> +	ret = gettimeofday(&tv_end, NULL);
> +	BUG_ON(ret);
>
>  	timersub(&tv_end, &tv_start, &tv_diff);
>
> diff --git a/tools/perf/bench/mem-memset.c b/tools/perf/bench/mem-memset.c
> index 350cc95..e0702d2 100644
> --- a/tools/perf/bench/mem-memset.c
> +++ b/tools/perf/bench/mem-memset.c
> @@ -139,17 +139,19 @@ static double do_memset_gettimeofday(memset_t
> fn, size_t len, bool prefault)
>  {
>  	struct timeval tv_start, tv_end, tv_diff;
>  	void *dst = NULL;
> -	int i;
> +	int i, ret;
>
>  	alloc_mem(&dst, len);
>
>  	if (prefault)
>  		fn(dst, -1, len);
>
> -	BUG_ON(gettimeofday(&tv_start, NULL));
> +	ret = gettimeofday(&tv_start, NULL);
> +	BUG_ON(ret);
>  	for (i = 0; i < iterations; ++i)
>  		fn(dst, i, len);
> -	BUG_ON(gettimeofday(&tv_end, NULL));
> +	ret = gettimeofday(&tv_end, NULL);
> +	BUG_ON(ret);

Ditto.


>
>  	timersub(&tv_end, &tv_start, &tv_diff);
>
> diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
> index 0c7454f..b35c94b 100644
> --- a/tools/perf/bench/sched-pipe.c
> +++ b/tools/perf/bench/sched-pipe.c
> @@ -61,8 +61,10 @@ int bench_sched_pipe(int argc, const char **argv,
>  	argc = parse_options(argc, argv, options,
>  			     bench_sched_pipe_usage, 0);
>
> -	assert(!pipe(pipe_1));
> -	assert(!pipe(pipe_2));
> +	ret = pipe(pipe_1);
> +	assert(!ret);
> +	ret = !pipe(pipe_2);
> +	assert(!ret);

What about converting these raw assert's to BUG_ONs (with negating the
conditions) for consistency?

Thanks,
Namhyung


>
>  	pid = fork();
>  	assert(pid >= 0);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/