Re: [PATCH 2/2] perf trace: Fix segfault on perf trace -i perf.data

[Namhyung Kim <namhyung@kernel.org>]

Hi Arnaldo,

On Tue, Nov 12, 2013 at 08:57:00AM -0300, Arnaldo Carvalho de Melo wrote:
> So this becomes the first part of this patch, split from yours and
> massaged a bit so that by looking at the patch it becomes quickly clear
> what it is doing, please let me now if I can keep this as-is (with your
> authorship, etc).

Looks good to me.

But I just have a nitpick, please see below.

> 
> I'll test this all out after finishing the next part of the split up.
> 
> commit 296f6ce34590099740bfe03ced37f6f53a0133f8
> Author: Namhyung Kim <namhyung@kernel.org>
> Date:   Tue Nov 12 08:51:45 2013 -0300
> 
>     perf trace: Separate tp syscall field caching into init routine to be reused
>     
>     We need to set this in evsels coming out of a perf.data file header, not
>     just for new ones created for live sessions.
> 
>     So separate the code that caches the syscall entry/exit tracepoint
>     format fields into a new function that will be used in the next
>     changeset.
>     
>     Signed-off-by: Namhyung Kim <namhyung@kernel.org>
>     Cc: Adrian Hunter <adrian.hunter@intel.com>
>     Cc: David Ahern <dsahern@gmail.com>
>     Cc: Frederic Weisbecker <fweisbec@gmail.com>
>     Cc: Jiri Olsa <jolsa@redhat.com>
>     Cc: Mike Galbraith <efault@gmx.de>
>     Cc: Paul Mackerras <paulus@samba.org>
>     Cc: Peter Zijlstra <peterz@infradead.org>
>     Cc: Stephane Eranian <eranian@google.com>
>     Link: http://lkml.kernel.org/n/tip-iv4vbx2064hc2drv38egqzee@git.kernel.org
>     Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> 
> diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
> index aeb6296a76bd..3fa1dce6d43e 100644
> --- a/tools/perf/builtin-trace.c
> +++ b/tools/perf/builtin-trace.c
> @@ -149,20 +149,32 @@ static void perf_evsel__delete_priv(struct perf_evsel *evsel)
>  	perf_evsel__delete(evsel);
>  }
>  
> +static int perf_evsel__init_syscall_tp(struct perf_evsel *evsel, void *handler)
> +{
> +	evsel->priv = malloc(sizeof(struct syscall_tp));
> +	if (evsel->priv != NULL) {
> +		if (perf_evsel__init_sc_tp_uint_field(evsel, id))
> +			goto out_delete;
> +
> +		evsel->handler = handler;
> +		return 0;
> +	}
> +
> +	return -ENOMEM;
> +
> +out_delete:
> +	free(evsel->priv);
> +	evsel->priv = NULL;

Is this part needed?  I can see that perf_evsel__delete_priv() can do
it for you anyway.  Yes I know it's needed for my later change, but I
think we do it a bit differently.

And again, is perf_evsel__delete_priv() needed?  Isn't the ->priv is
not used for anything else?  Why not just letting perf_evsel__delete()
handle this transparently?

Looking at the source, evsel->priv is a member of union and the other
member ->id_offset is used in when dealing with the perf file header
and it doesn't allocate memory.

Hmm, how about adding a new field like ->needs_free_priv then?

Anyway, it should definitely be a different change, I just want to
raise an issue after seeing it.

Thanks,
Namhyung


> +	return -ENOENT;
> +}
> +
>  static struct perf_evsel *perf_evsel__syscall_newtp(const char *direction, void *handler)
>  {
>  	struct perf_evsel *evsel = perf_evsel__newtp("raw_syscalls", direction);
>  
>  	if (evsel) {
> -		evsel->priv = malloc(sizeof(struct syscall_tp));
> -
> -		if (evsel->priv == NULL)
> +		if (perf_evsel__init_syscall_tp(evsel, handler))
>  			goto out_delete;
> -
> -		if (perf_evsel__init_sc_tp_uint_field(evsel, id))
> -			goto out_delete;
> -
> -		evsel->handler = handler;
>  	}
>  
>  	return evsel;