Closed staticfloat closed 2 years ago
Is it reproducible with hardware breakpoints (i.e. hbreak *0xfoo
instead of b *0xfoo
)?
No, using a hardware-assisted breakpoint seems to work:
(rr) hbreak *0x7f4246aff6f0
Hardware assisted breakpoint 1 at 0x7f4246aff6f0
(rr) rc
Continuing.
Thread 2 received signal SIGSEGV, Segmentation fault.
0x00007f424735135e in ?? ()
(rr) rc
Continuing.
Idk if @Keno wants to dig in here but something reading the executable code and noticing the breakpoint would explain this and we've seen it before in other contexts.
If hbreak
works, probably not worth debugging.
I am using
rr
to debug a segfault in Julia when working with a USB device. I have noticed that a certain order of commands reliably causes a replay divergence. The commands are:`rr` output
The output of running `rr replay`: ``` $ ~/src/rr/bin/rr replay . GNU gdb (Ubuntu 12.0.90-0ubuntu1) 12.0.90 Copyright (C) 2022 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or laterThe packed trace is available here. The machine this was run on is an AMD machine:
The
rr
version used was the latestmaster
as of this writing: