Re: [PATCH] perf script: Fix overrun issue for dynamically-allocated pmu type number

[Adrian Hunter <adrian.hunter@intel.com>]

On 26/11/20 9:06 am, Jin, Yao wrote:
> Hi Adrian,
> 
> On 11/26/2020 2:51 PM, Adrian Hunter wrote:
>> On 26/11/20 5:24 am, Jin Yao wrote:
>>> When unpacking the event which is from dynamic pmu, the array
>>> output[OUTPUT_TYPE_MAX] may be overrun. For example, type number of
>>> SKL uncore_imc is 10, but OUTPUT_TYPE_MAX is 7 now (OUTPUT_TYPE_MAX =
>>> PERF_TYPE_MAX + 1).
>>>
>>> /* In builtin-script.c */
>>> process_event()
>>> {
>>>     unsigned int type = output_type(attr->type);
>>>
>>>     if (output[type].fields == 0)
>>>         return;
>>> }
>>>
>>> output[10] is overrun.
>>>
>>> Create a type OUTPUT_TYPE_OTHER for dynamic pmu events, then
>>> output_type(attr->type) will return OUTPUT_TYPE_OTHER here.
>>>
>>> Note that if PERF_TYPE_MAX ever changed, then there would be a conflict
>>> between old perf.data files that had a dynamicaliy allocated PMU number
>>> that would then be the same as a fixed PERF_TYPE.
>>>
>>> Example:
>>>
>>> perf record --switch-events -C 0 -e
>>> "{cpu-clock,uncore_imc/data_reads/,uncore_imc/data_writes/}:SD" -a --
>>> sleep 1
>>> perf script
>>>
>>> Before:
>>>           swapper     0 [000] 1479253.987551:     277766              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1479253.987797:     246709              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1479253.988127:     329883              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1479253.988273:     146393              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1479253.988523:     249977              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1479253.988877:     354090              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1479253.989023:     145940              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1479253.989383:     359856              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1479253.989523:     140082              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>
>>> After:
>>>           swapper     0 [000] 1397040.402011:     272384              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1397040.402011:       5396 
>>> uncore_imc/data_reads/:
>>>           swapper     0 [000] 1397040.402011:        967
>>> uncore_imc/data_writes/:
>>>           swapper     0 [000] 1397040.402259:     249153              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1397040.402259:       7231 
>>> uncore_imc/data_reads/:
>>>           swapper     0 [000] 1397040.402259:       1297
>>> uncore_imc/data_writes/:
>>>           swapper     0 [000] 1397040.402508:     249108              
>>> cpu-clock:  ffffffff9d4ddb6f cpuidle_enter_state+0xdf ([kernel.kallsyms])
>>>           swapper     0 [000] 1397040.402508:       5333 
>>> uncore_imc/data_reads/:
>>>           swapper     0 [000] 1397040.402508:       1008
>>> uncore_imc/data_writes/:
>>>
>>> Fixes: 1405720d4f26 ("perf script: Add 'synth' event type for synthesized
>>> events")
>>
>> It does not look to me like the problem was introduced by that commit.  Are
>> you sure this Fixes tag is correct?
>>
> 
> Commit 1405720d4f26 added the change:
> 
> @@ -1215,8 +1253,9 @@ static void process_event(struct perf_script *script,
>  {
>         struct thread *thread = al->thread;
>         struct perf_event_attr *attr = &evsel->attr;
> +       unsigned int type = output_type(attr->type);
> 
> -       if (output[attr->type].fields == 0)
> +       if (output[type].fields == 0)
>                 return;

That is a nop if attr->type != PERF_TYPE_SYNTH
Given that PERF_TYPE_SYNTH is (INT_MAX + 1), it is a nop for all kernel
dynamically allocated PMU numbers.

> 
> But of course, we can also say the original "output[attr->type].fields"
> introduced the issue, I'm not sure. Maybe Arnaldo can help to make the
> decision. :)

I think perf script has always had this problem.