00:11 <agentzh> fche: i see. thanks

00:13 <agentzh> kerneltoast: so better check the staprun output, if they give the expected output as normal runs. as fche mentioned, some of the probe process.begin handlers may run on unwanted processes. we only care about our own target processes, which are just a few or just a single one.

00:14 <agentzh> now the consensus is just to ignore those errors instead of making them fatal?

00:38 <fche> I think they never were fatal

00:38 <fche> in the sense of stopping the stap script

00:38 <fche> more like warnings that some probes were not attached as intended (probably)

15:16 <modem> hi

15:16 <modem> i am using a lot of absolute probe

15:16 <modem> but i need to manually edit after each reboot because of base_address kernel KASLR

15:17 <modem> I guess it could be automated easily compute the proper offset at least when the probe cannot be inserted (or with a new getopt switch)

15:17 <modem> i can offer you to do it on my own and git pull

15:17 <modem> any hints/advice on what to do is welcome. Thanks ! :-)

21:29 <agentzh> modem: maybe disabling kaslr is the easiest approach?

21:30 <agentzh> or you can generate your .stp script with another perl/python/ruby script which finds the offsets?

23:12 <fche> can you remind me why you're using absolute probes so much, vs. normal symbolic probes, or kprobe.function("xxx") ?