Hi Oleg,
> For the moment, please forget about ret-probes. Could you compile this program
>
> #define _GNU_SOURCE
> #include <unistd.h>
> #include <sched.h>
> #include <signal.h>
>
> int func(int i)
> {
> return i;
> }
>
> int test(void *arg)
> {
> int i;
> for (i = 0;; ++i) {
> sleep(1);
> func(i);
> }
> return 0;
> }
>
> int main(void)
> {
> static char stack[65536];
>
> clone(test, stack + sizeof(stack)/2, CLONE_VM|SIGCHLD, NULL);
> test(NULL);
>
> return 0;
> }
>
> and then do something like
>
> $ ./test &
> $ bpftrace -p $! -e 'uprobe:./test:func { printf("%d\n", pid); }'
>
> I hope that the syntax of the 2nd command is correct...
>
> I _think_ that it will print 2 pids too.
>
> But "perf-record -p" works as expected.
Yes, the output from bpftrace and perf matches what you described:
$ ./tester &
[1] 158592
$ perf probe -x tester --add func
Added new event:
probe_tester:func (on func in /root/test/tester)
$ bpftrace -p 158592 -e 'uprobe:./tester:func { printf("time=%llu pid=%d\n", elapsed / 1000000000, pid); }'
Attaching 1 probe...
time=0 pid=158592
time=0 pid=158594
time=1 pid=158592
time=1 pid=158594
time=2 pid=158592
time=2 pid=158594
time=3 pid=158592
time=3 pid=158594
$ perf record -e probe_tester:func -p 158592 -o 158592
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.022 MB 158592 (19 samples) ]
$ perf script -i 158592
tester 158592 [006] 246475.295762: probe_tester:func: (55a6def14149)
tester 158592 [006] 246476.295828: probe_tester:func: (55a6def14149)
tester 158592 [006] 246477.295892: probe_tester:func: (55a6def14149)
tester 158592 [006] 246478.295958: probe_tester:func: (55a6def14149)
tester 158592 [006] 246479.296024: probe_tester:func: (55a6def14149)
tester 158592 [010] 246480.296202: probe_tester:func: (55a6def14149)
tester 158592 [010] 246481.296360: probe_tester:func: (55a6def14149)
[...]
Thanks,
