On Tue, Jun 24, 2025 at 04:44:49PM +0200, Gabriele Monaco wrote:

As you mentioned, nobody is likely relying on those tracepoints names
at the moment, but I would rather be cautious basing userspace tools on
some monitors to exist at all.

In my opinion, RV tracepoints are useful as an alternative of
reactors/rv userspace tool but cannot be used without considering the
RV interface itself (e.g. available_monitors and friends).

We could at least stick to the following assumptions:
1. monitors can change names, be added or removed
2. tracepoints are consistent to monitor names (in available_monitors)
3. the tracepoint structure does not change (i.e. event_/error_, args.)
   (can change for new monitors types where seen fit)

If in the future we allow the possibility to build RV monitors as BPF
programs, we'd probably also allow monitors without tracepoints at all,
but I'd say for now those assumptions are sensible.

What do you think?

I would like that. Ideally, the userspace tools only use tracepoints based
on available_monitors.

However, people may not do that, and just use tracepoints directly.

You could argue that those tools are not correctly designed. Therefore it
is their fault that the tools are broken after updating kernel.

On the other hand, there is this sentiment that we must never break
userspace.

I don't know enough to judge this. Maybe @Steven has something to add?

Best regards,
Nam