RE: [bitbake-devel] [PATCH] utils: Enable the loopback interface in disable_network()

[Peter Kjellerstedt <peter.kjellerstedt@axis.com>]

> -----Original Message-----
> From: Richard Purdie <richard.purdie@linuxfoundation.org>
> Sent: den 23 september 2022 16:54
> To: Peter Kjellerstedt <peter.kjellerstedt@axis.com>; bitbake-
> devel@lists.openembedded.org
> Subject: Re: [bitbake-devel] [PATCH] utils: Enable the loopback interface
> in disable_network()
> 
> On Fri, 2022-09-23 at 14:42 +0000, Peter Kjellerstedt wrote:
> > Ok. I hope it is then ok that I add a new task varflag "network-loopback"
> > (or maybe even just "loopback"?) E.g.:
> >
> > diff --git a/bitbake/bin/bitbake-worker b/bitbake/bin/bitbake-worker
> > index 7be39370b3..4115604d99 100755
> > --- a/bitbake/bin/bitbake-worker
> > +++ b/bitbake/bin/bitbake-worker
> > @@ -267,6 +267,8 @@ def fork_off_task(cfg, data, databuilder,
> > workerdata, fn, task, taskname, taskha
> >                      if bb.utils.is_local_uid(uid):
> >                          logger.debug("Attempting to disable network for %s" % taskname)
> >                          bb.utils.disable_network(uid, gid)
> > +                        if the_data.getVarFlag(taskname, 'network-loopback', False):
> > +                            bb.utils.loopback_up()
> >                      else:
> >                          logger.debug("Skipping disable network for %s since %s is not a local uid." % (taskname, uid))
> >
> > With that in place, I see no drawbacks compared to always
> > enabling the loopback interface in disable_network() since
> > setting, e.g., do_compile[network-loopback] = "1" will make
> > sure it is only enabled when actually needed, compared to,
> > e.g., doing it in an anonymous Python function.
> 
> Setting it in a prefunc doesn't look that bad once you have a function
> for it?
> 
> python enable_loopback () {
>     bb.utils.loopback_up()
> }
> 
> do_compile[prefuncs] += "enable_loopback"

Hmm. What worries me here is what happens when enable_loopback is 
called for a task that has network=1? Hopefully it only does some 
unnecessary work, but I do not know... That's why I find the 
varflag solution more appealing as I know it will only enable the 
loopback interface when networking has actually been disabled.

> On the subject of naming, it may be clearer as something like:
> 
> bb.utils.enable_loopback_networking()
> 
> as we need to think about how this looks without context we have in
> these patches.

Agreed.

> Cheers,
> 
> Richard

//Peter