Re: [PATCH] perf: fix a segfault problem.

[Wang Nan <wangnan0@huawei.com>]

On 2015/3/13 17:46, Namhyung Kim wrote:
> On Fri, Mar 13, 2015 at 08:41:32AM +0000, Wang Nan wrote:
>> Without this patch, perf report cause segfault if pass "" as '-t':
>>
>>   $ perf report -t ""
>>
>>     # To display the perf.data header info, please use --header/--header-only options.
>>     #
>>     # Samples: 37  of event 'syscalls:sys_enter_write'
>>     # Event count (approx.): 37
>>     #
>>     # Children    SelfCommand   Shared Object         Symbol
>>     Segmentation fault
>>
>> This patch avoid the segfault by checking empty string for
>> 'symbol_conf.field_sep'.
> 
> What about resetting it to NULL if empty string was given?
> 

In fact I'm not very clear why we need such 'symbol_conf.field_sep', so I'm
not sure whether '-t ""' is totally meanless or not.

-t option replaces a group of character with '.' and appends them after a field.
With -t 'abc' I get something like:

 #
 # OverheadabcCommand   abcShared Object    abcSymbol
  100.00%abcb.beltr.ceabc[kernel.k.llsyms]abc[k] 0xffffffff810118f0
  ...

Hard to read...

I read docs and your commit messages, but still not understand the option. Could you
please explain the goal and usage of that option again?

Thank you.

> Thanks,
> Namhyung
> 
> 
>>
>> Signed-off-by: Wang Nan <wangnan0@huawei.com>
>> ---
>>  tools/perf/util/sort.c | 3 ++-
>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
>> index 4593f36..7f563a0 100644
>> --- a/tools/perf/util/sort.c
>> +++ b/tools/perf/util/sort.c
>> @@ -31,7 +31,8 @@ static int repsep_snprintf(char *bf, size_t size, const char *fmt, ...)
>>  
>>  	va_start(ap, fmt);
>>  	n = vsnprintf(bf, size, fmt, ap);
>> -	if (symbol_conf.field_sep && n > 0) {
>> +	if (symbol_conf.field_sep && n > 0 &&
>> +			(symbol_conf.field_sep[0] != '\0')) {
>>  		char *sep = bf;
>>  
>>  		while (1) {
>> -- 
>> 1.8.3.4
>>