On 08/24, Tianyi Liu wrote:

A key trigger here is that the two tracer instances (either `bpftrace` or
`perf record`) must be running *simultaneously*. One of them should use
PID1 as filter, while the other uses PID2.

Yes.

Agreed, the implementation of uretprobe should be almost the same as
uprobe, but it seems uretprobe was ignored in previous modifications.

I forgot EVERYTHING about the code in kernel/trace/trace_uprobe.c, but
at first glance I am not sure that your patch is what we need...

At least I certainly disagree with "Fixes: c1ae5c75e103" ;)

uretprobe_perf_func/etc was designed for perf, and afaics this code still
works fine even if you run 2 perf-record's with -p PID1/PID2 at the same
time.

BPF hacks/hooks were added later, so perhaps this should be fixed in the
bpf code, but I have no idea what bpftrace does...

Oleg.