Re: [PATCH v5] ALSA: control: Verify put() result when in debug mode

[Cezary Rojewski <cezary.rojewski@intel.com>]

On 2026-02-19 11:30 AM, Takashi Iwai wrote:
> On Thu, 19 Feb 2026 11:15:41 +0100,
> Cezary Rojewski wrote:

...

>> In regard to the "trace system", not a fan of adding a new "system"
>> just for the sake of a single event.
> 
> Why not?  We may extend the tracing of controls, too, and it'll make
> easier to enable all tracepoints relevant either for PCM or control.
> 
>> If we are going that route, I'd
>> rather reword "snd_pcm" to "snd" or something similar. From sound-core
>> perspective, I believe there should be 1 directory for the events
>> declared in sound/core. The event-names are prefixed already, further
>> division looks redundant.
> 
> That'll be messy under /sys/kernel/debug/tracing/events/*; it'll
> become snd/applptr, snd/hwptr, etc.  So, if we want to consolidate
> into a single directory, we need to put "pcm_" prefix again to each.
> 
> Or maybe better to use subdirectories "snd/pcm" and "snd/ctl" instead.

Well, I'm taking example from existing directories e.g.: 
trace/events/power or trace/events/sched. TBH, trace/event/asoc fits 
here too - all of them just group events together. sched_kthread_xxx are 
in the same bucket as sched_process_xxx (as are plenty of others). The 
'name' alone is enough of a filter. I believe that until one reaches 
certain number of event-types, there is no reason to dilute the hierarchy.

However, looking at the subject from a high-level perspective, I think 
it isn't in the scope of this patch - I can align to any suggested 
pattern and we can work on the details in a dedicated series.


Kind regards,
Czakre