Re: [PATCH] tracing: Add __cpumask to denote a trace event field that is a cpumask_t

[Valentin Schneider]

On 15/10/22 07:14, Steven Rostedt wrote:
> On Fri, 14 Oct 2022 18:21:41 +0100
> Valentin Schneider <vschneid@xxxxxxxxxx> wrote:
>> 
>> Thanks for spinning this out so quickly! I gave it a test against my IPI
>> tracepoints, it's working OK with one (expected) change:
>
> No problem. It was my "While traveling I pick what I want to work on" project ;-)
>
>> 
>>   [ipi:ipi_send_cpumask] function __get_cpumask not defined
>>   CPU 0 is empty
>>   CPU 1 is empty
>>   CPU 3 is empty
>>   cpus=4
>>             echo-173   [002]    11.859745: ipi_send_cpumask:     [FAILED TO PARSE] cpumask=ARRAY[02, 00, 00, 00, 00, 00, 00, 00] callsite=0xffffffff81121013
>> 
>> so libtraceevent is going to need updating - I'm happy to do that (if you
>> haven't done it already!)
>
> I have not done that. I'm happy to take a patch from you.
>

I'll go do that and try to mess around the pretty-print output as well.

>> 
>> Lastly, given all cpumasks have a (usable) size of nr_cpumask_bits, we can
>> factor out the size argument, see below. Regardless:
>
> Seems reasonable. I can fold that into my patch and if you reply back
> with your signed-off-by tag, I'll include you as co-author.
>