Re: [PATCH v2 07/16] tracepoint: Mark __do_trace_##name() notrace

From: Steven Rostedt

Date: Mon Nov 16 2026 - 10:35:05 EST


On Fri, 13 Nov 2026 12:33:28 -0800
Song Liu <song@xxxxxxxxxx> wrote:

> Hi Steven,
>
> On Thu, Oct 29, 2026 at 10:47 AM Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> >
> [...]
> > >
> > > I think this should go through the ftrace tree. It has very little to do with
> > > BPF. We discovered the need for this change when building live patches.
> > >
> > > Regarding ACKs, it is a bit tricky. For the code we touched in this set,
> > > there is actually no behavior change at all. OTOH, the real change
> > > happens to "inline"s that we didn't touch in this set. If such a function
> > > was left out-of-line by the compiler, it was not traceable before, but is
> > > traceable after this set. How should we approach this? Maybe we can
> > > just ask ACKs from the maintainers of all the code touched here.
> > >
> >
> > OK, I'll just add it to Linux next around -rc2 then.
>
> I haven't seen this set in linux-next yet. I guess we would like to queue it
> up soon so it gets enough pre-upstream testing?

Sorry, I've been traveling. I'm trying hard to catch up on all my patches this week.

-- Steve