On Wed, 2026-10-21 at 13:12 +0200, Nam Cao wrote: > Gabriele Monaco <[email protected]> writes: > > On Wed, 2026-10-21 at 10:13 +0200, Nam Cao wrote: > > > I wouldn't enable those by default though, it's way too much noises. As > > > it's for debugging the model, we can have some sort of compile-time > > > switch to turn it on. > > > > That's almost what's happening already, in-kernel use tracepoints, gated by > > static branches. Those are the closest thing you can get to a dynamic > > compile- > > time switch (assuming such a thing can exist), it's patching the code. > > I was talking about how people are going to enable transition events, > not how kernel enables them. Do we expose some user API for people to > enable at run-time (and kernel can use run-time patching or whatever), > or people enable them while compiling their BPF programs. I think the > latter should suffice.
Everything BPF-related will need to go through the userspace tool. From loading monitors to enabling things, but it shouldn't require any recompilation. Right now I use only the -t flag (enable tracing and print to stdout), we'd need some flag to only enable tracing. After that you should be able to use ftrace as normal. Gabriele
