Comment 3 for bug 1838575

> > sudo perf record -a -F $((250*100)) -e cycles:u --
> > /usr/bin/qemu-system-x86_64 -name guest="guest" ....
>
> I instead attached perf to the qemu process after it was spawned by
> libvirt so I didn't have to worry about producing a working qemu
> cmdline. I let it run for several seconds while still in the paused
> state.

Sure!

> > And check where we're spending so much time for the boot to happen.
> > Note I'm using "cycles:u", getting only userland stack trace probes
> > every (HZ*100),.
>
> This showed no samples - apparently all samples were kernel side.

Either kernel or VMX/SVM mode.

> > That might give us enough information.
> >
> > Would you mind opening a launchpad bug (can be private if you prefer)
> > so Christian and I can take a look ?
>
> Sure:
> https://bugs.launchpad.net/ubuntu/+source/qemu/+bug/1838575

Err, sorry to ask you this, but I'm getting magic/endian check errors
on perf report, maybe something related to the arch and perf. Would
you mind reporting to stdio and pasting into the case ? Are you using
fully updated Eoan (userland/kernel) for everything ?

(c)inaddy@temp:~/temp$ sudo /usr/lib/linux-tools/5.2.0-8-generic/perf
report -f -v
magic/endian check failed

Tks!