Re: [PATCH v5] block: Add ioprio to block_rq tracepoint

[Bart Van Assche <bvanassche@acm.org>]

On 6/17/24 12:59 AM, dongliang cui wrote:
> On Sat, Jun 15, 2024 at 12:41 AM Bart Van Assche <bvanassche@acm.org> wrote:
>>
>> On 6/14/24 12:49 AM, Dongliang Cui wrote:
>>> -     TP_printk("%d,%d %s (%s) %llu + %u [%d]",
>>> +     TP_printk("%d,%d %s (%s) %llu + %u %s,%u,%u [%d]",
>>>                  MAJOR(__entry->dev), MINOR(__entry->dev),
>>>                  __entry->rwbs, __get_str(cmd),
>>> -               (unsigned long long)__entry->sector,
>>> -               __entry->nr_sector, 0)
>>> +               (unsigned long long)__entry->sector, __entry->nr_sector,
>>> +               __print_symbolic(IOPRIO_PRIO_CLASS(__entry->ioprio),
>>> +                                IOPRIO_CLASS_STRINGS),
>>> +               IOPRIO_PRIO_HINT(__entry->ioprio),
>>> +               IOPRIO_PRIO_LEVEL(__entry->ioprio),  0)
>>>    );
>>
>> Do we really want to include the constant "[0]" in the tracing output?
> This is how it is printed in the source code.
>  From the code flow point of view, there is no need to print this value
> in trace_block_rq_requeue.
> Do we need to consider the issue of uniform printing format? If not, I
> think we can delete it.

I'm not aware of any other tracing statement that prints out a constant.
Is there perhaps something that I'm missing or overlooking?

Thanks,

Bart.