Re: [PATCH v2 1/2] tracing/hist: rebuild full_name on each hist_field_name() call

[Steven Rostedt <rostedt@goodmis.org>]

On Wed, 8 Apr 2026 10:15:00 +0800
Pengpeng Hou <pengpeng@iscas.ac.cn> wrote:

> Hi Steve,
> 
> Thanks for looking at this.
> 
> I believe this path is reachable.
> 
> field->system is set when a VAR_REF is created with an event-qualified
> reference. The assignment happens in init_var_ref(), which is called
> from create_var_ref().
> 
> The parser accepts fully qualified references of the form
> 
>   system.event.$var
> 
> and parse_expr() passes the parsed system/event strings into
> create_var_ref(). trace_action_create() also accepts explicitly
> qualified action parameters in the same form and passes those strings
> into create_var_ref().
> 
> This is also documented in Documentation/trace/histogram.rst:
> 
>   fully-qualified name is of the form 'system.event_name.$var_name'
>   or 'system.event_name.field'.

Ah I read the document and missed this part. Thanks for pointing it out.

None of our tests use this, so that definitely needs to be fixed.

OK, yeah I can trigger that path now. I'll continue looking at the patch.

Thanks,

-- Steve


> 
> The path is user-visible because reading back the trigger goes through
> 
>   event_hist_trigger_print() -> hist_field_print() -> hist_field_name()
> 
> so creating a hist trigger that uses an inter-event variable reference
> in fully qualified form is enough to exercise it.
> 
> So I don't think this is dead code. If you would prefer a different way
> of handling the printed name, I can respin accordingly, but I don't
> think the current branch is unreachable.
> 
> Thanks,
> Pengpeng
>