Paolo Bonzini <pbonz...@redhat.com> writes: > On 14/03/2017 11:05, Paolo Bonzini wrote: >> >> >> On 13/03/2017 19:15, Alex Bennée wrote: >>>> Well yeah, it's all those that pass qemu_clock_use_for_deadline. >>> >>> Have you done any testing with record/replay? So far I have one >>> reproducible run and one failure. However it is not entirely clear to me >>> how I am meant to cleanly halt and stop a machine so I don't corrupt the >>> replay log. >> >> No, I haven't tried RR. > > Tried now, it doesn't work but I won't have time to fix it. > > I tried "-kernel /path/to/bzImage -icount 3,rr=record,rrfile=replay.bin > -net none". record works, replay hangs. The hang is at a "pause" > instruction. With "-d in_asm" it still hangs, but a bit later (still in > the firmware).
I'm happy to keep digging. So to be clear when you do the record step how do you save your state before you replay? I was just entering the MMI and typing quit but that seems to leave the log in a broken state. -- Alex Bennée
