Re: [PATCH v8 13/21] main: keep rcu_atfork callback enabled for qtest

[Alexander Bulekov]

On 200618 0934, Thomas Huth wrote:
> On 29/01/2020 06.34, Bulekov, Alexander wrote:
> > The qtest-based fuzzer makes use of forking to reset-state between
> > tests. Keep the callback enabled, so the call_rcu thread gets created
> > within the child process.
> > 
> > Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
> > ---
> >  vl.c | 9 ++++++++-
> >  1 file changed, 8 insertions(+), 1 deletion(-)
> > 
> > diff --git a/vl.c b/vl.c
> > index bb77935f04..cf8e2d3ebb 100644
> > --- a/vl.c
> > +++ b/vl.c
> > @@ -3794,7 +3794,14 @@ void qemu_init(int argc, char **argv, char **envp)
> >      set_memory_options(&ram_slots, &maxram_size, machine_class);
> >  
> >      os_daemonize();
> > -    rcu_disable_atfork();
> > +
> > +    /*
> > +     * If QTest is enabled, keep the rcu_atfork enabled, since system 
> > processes
> > +     * may be forked testing purposes (e.g. fork-server based fuzzing)
> > +     */
> > +    if (!qtest_enabled()) {
> > +        rcu_disable_atfork();
> > +    }
> >  
> >      if (pid_file && !qemu_write_pidfile(pid_file, &err)) {
> >          error_reportf_err(err, "cannot create PID file: ");
> 
>  Hi Alexander,
> 
> I think this patch might maybe not work as expected: The qtest_enabled()
> has been added before configure_accelerators() is called in main(). So
> qtest_enabled() should always return "false" and thus
> rcu_disabled_fork() is still called in any case... could you please
> double-check whether it works for you and I just made a mistake, or
> whether this is a bug indeed?
Hi,
This is a problem.. I think we can work around this by calling
rcu_enable_atfork from the fuzzer, after qemu_init. I'll send a patch
soon.
Thanks
-Alex

>  Thanks,
>   Thomas
>