Re: [PATCH v3 4/4] trace/preemptirq: Implement trace_irqflags hooks

[Peter Zijlstra <peterz@infradead.org>]

On Wed, Mar 11, 2026 at 03:48:42PM -0400, Steven Rostedt wrote:
> On Wed, 11 Mar 2026 20:43:05 +0100
> Peter Zijlstra <peterz@infradead.org> wrote:
> 
> > On Wed, Mar 11, 2026 at 09:50:18AM -0300, Wander Lairson Costa wrote:
> > > +#define local_irq_enable()				\
> > > +	do {						\
> > > +		if (tracepoint_enabled(irq_enable))	\
> > > +			trace_local_irq_enable();	\  
> > 
> > I'm thinking you didn't even look at the assembly generated :/
> > 
> > Otherwise you would have written this like:
> > 
> > 		if (tracepoint_enabled(irq_enable))
> > 			__do_trace_local_irq_enable();
> 
> Please don't use the internal functions outside of the tracepoint.h
> 
> Vineeth is currently working on a patch set to properly do that. It's going
> to introduce:
> 
>   trace_invoke_<event>()
> 
> Which basically is just __do_trace_<event>(), but as a wrapper that can
> handle updates that may be needed, but supplies a proper API where thing
> wont randomly break when __do_trace_<event>() changes.

That's like a 3 line patch, hardly worth the effort. Its not like it'll
be hard to find and fix any users if you do ever change that.