linux-kernel - Re: [PATCH 00/15] tools: Unify perf and trace-cmd trace event format parsing v3

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]

Date:	Mon, 7 May 2012 16:13:31 +0200
From:	Ingo Molnar <mingo@...nel.org>
To:	Steven Rostedt <rostedt@...dmis.org>
Cc:	Frederic Weisbecker <fweisbec@...il.com>,
	LKML <linux-kernel@...r.kernel.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Peter Zijlstra <peterz@...radead.org>,
	Arnaldo Carvalho de Melo <acme@...radead.org>,
	Borislav Petkov <bp@...en8.de>, Jiri Olsa <jolsa@...hat.com>,
	Arun Sharma <asharma@...com>,
	Namhyung Kim <namhyung.kim@....com>,
	Michael Rubin <mrubin@...gle.com>,
	David Sharp <dhsharp@...gle.com>,
	Vaibhav Nagarnaik <vnagarnaik@...gle.com>,
	Julia Lawall <julia@...u.dk>, Tom Zanussi <tzanussi@...il.com>,
	David Ahern <dsahern@...il.com>
Subject: Re: [PATCH 00/15] tools: Unify perf and trace-cmd trace event format
 parsing v3


* Steven Rostedt <rostedt@...dmis.org> wrote:

> On Mon, 2012-05-07 at 10:14 +0200, Ingo Molnar wrote:
> 
> > So can we please make a libevent.so, built sanely within 
> > tools/perf/lib/ or such and distributed together with perf so 
> > that the two can never get out of sync?
> 
> If you do this, you need to find a way to turn off -Wswitch-enum as that
> seems to (at least on my gcc) ignore defaults.
> 
> Thus we end up with:
> 
> diff --git a/Makefile b/Makefile
> index 18ad079..217548b 100644
> 
> diff --git a/parse-filter.c b/parse-filter.c
> index d09fbd2..0ac249c 100644
> --- a/parse-filter.c
> +++ b/parse-filter.c
> @@ -367,6 +367,12 @@ create_arg_item(struct event_format *event, const
> char *token,
>  		arg->type = FILTER_ARG_FIELD;
>  		arg->field.field = field;
>  		break;
> +	case EVENT_ERROR:
> +	case EVENT_NONE:
> +	case EVENT_SPACE:
> +	case EVENT_NEWLINE:
> +	case EVENT_OP:
> +	case EVENT_DELIM:
>  	default:
>  		free_arg(arg);
>  		show_error(error_str, "expected a value but found %s",

The advantage of this warning is that when a new enum value is 
added, every usage site is forced to consider it.

Code not using it indeed needs to be updated.

> Looking at what was done in the current code, there's lots of :
> 
> 	case PRINT_NULL:
> 	case PRINT_FIELD ... PRINT_SYMBOL:
> 	case PRINT_STRING:
> 	default:
> 
> Which looks more of a maintenance nightmare. How does this help? The
> FOO ... BAR, will hide the same errors that you are trying to prevent.

Indeed the FOO...BAR pattern should not be used (i.e. the above 
is a bug in essence), unless it's clearly some genuine same-type 
numeric range that is expressed, where new fields can never get 
inbetween.

> If you were suppose to handle something between FOO and BAR, then you
> just ignored it too.

Correct.

> It also makes that "default" redundant.

Yes.

> This is one of those warnings that causes more pain than it 
> helps.

I disagree, when I switched to it then it found some real bugs 
and new changes were incremental.

> If you strongly believe that all these warnings are helpful, 
> than we should push to add these warnings to the kernel too.

For subsystems I maintain we could consider it - but for the 
kernel as a whole it would be quite some work.

Thanks,

	Ingo
--
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/