Re: Wrong Perf Backtraces

Thanks.
I used an evince-3.28.4 and the following command:
sudo perf record -d --call-graph dwarf -c 100 -e 
mem_load_uops_retired.l3_miss:uppp /opt/evince-3.28.4/bin/evince

I only opened a file and, then, closed it using ALT+F4. The repeated 
gmallocn()s is reproducible with any sampling frequency, but the other 
backtrace is sometimes not reproducible in low frequencies. Where should 
I send you the log file?
Regards.

On 2020-03-25 20:16, Jiri Olsa wrote:

> On Wed, Mar 25, 2020 at 07:48:39PM +0430, ahmadkhorrami wrote:
> 
>> Hi,
>> 
>> Could you give me some hints about where the actual problem takes 
>> place? Is
>> the problem with "Perf" or the hardware part (i.e., "Hardware 
>> Performance
>> Counters")? Can I revise the problem by simply modifying the code? How 
>> much
>> work is needed?
> 
> heya,
> might be some callchain processing bug, but I can't reproduce it on my 
> setup..
> would you have/make some simple example that would reproduce the issue?
> 
> Another option is that you'd send perf.data together with 'perf 
> archive' data.
> 
> Also.. we support 2 dwarf unwinders (libunwind/libdw).. not sure which 
> one you
> have compiled in, but would be helpful to see if the other shows the 
> same.
> 
> jirka