On Tue, Sep 2, 2025 at 7:38 AM Jiri Olsa <jo...@kernel.org> wrote: > > Adding support to attach unique probe through perf uprobe pmu. > > Adding new 'unique' format attribute that allows to pass the > request to create unique uprobe the uprobe consumer. > > Signed-off-by: Jiri Olsa <jo...@kernel.org> > --- > include/linux/trace_events.h | 2 +- > kernel/events/core.c | 8 ++++++-- > kernel/trace/trace_event_perf.c | 4 ++-- > kernel/trace/trace_probe.h | 2 +- > kernel/trace/trace_uprobe.c | 9 +++++---- > 5 files changed, 15 insertions(+), 10 deletions(-) > > diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h > index 04307a19cde3..1d35727fda27 100644 > --- a/include/linux/trace_events.h > +++ b/include/linux/trace_events.h > @@ -877,7 +877,7 @@ extern int bpf_get_kprobe_info(const struct perf_event > *event, > #endif > #ifdef CONFIG_UPROBE_EVENTS > extern int perf_uprobe_init(struct perf_event *event, > - unsigned long ref_ctr_offset, bool is_retprobe); > + unsigned long ref_ctr_offset, bool is_retprobe, > bool is_unique);
In bpf land we don't allow multiple bool arguments any more. It makes callsites hard to read/review/maintain. Here I recommend to use enum flags as well.
