Arnaldo Carvalho de Melo wrote:
> Em Sat, Jul 31, 2010 at 08:27:48AM +0530, Srikar Dronamraju escreveu:
>>>> @@ -1598,15 +1812,19 @@ struct __event_package {
>>>>  int add_perf_probe_events(struct perf_probe_event *pevs, int npevs,
>>>>  			  bool force_add, int max_tevs)
>>>>  {
>>>> -	int i, j, ret;
>>>> +	int i, j, ret = 0;
>>>>  	struct __event_package *pkgs;
>>>> 
>>>>  	pkgs = zalloc(sizeof(struct __event_package) * npevs);
>>>>  	if (pkgs == NULL)
>>>>  		return -ENOMEM;
>>>> 
>>>> -	/* Init vmlinux path */
>>>> -	ret = init_vmlinux();
>>>> +	if (!pevs->upid)
>>>> +		/* Init vmlinux path */
>>>> +		ret = init_vmlinux();
>>>> +	else
>>>> +		ret = init_perf_uprobes();
>>>> +
>>>>  	if (ret < 0)
>>> pkgs leaks here.
>> Right, but I dont think this leak was introduced by my patch(s). I
>> guess its better fixed by a different patch. 
> 
> Sorry, will fix that then on a separate patch.

Yes, sorry, that was introduced by me...

Thanks for pointing it out.

-- 
Masami HIRAMATSU
2nd Research Dept.
Hitachi, Ltd., Systems Development Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com