Re: [PATCH 0/1] qom: fix setting of qdev array properties

["Daniel P. Berrangé" <berrange@redhat.com>]

On Fri, Sep 08, 2023 at 02:16:35PM +0200, Kevin Wolf wrote:
> Am 08.09.2023 um 11:27 hat Daniel P. Berrangé geschrieben:
> > On Fri, Sep 08, 2023 at 11:25:54AM +0200, Kevin Wolf wrote:
> > > Am 07.09.2023 um 11:35 hat Peter Maydell geschrieben:
> > > > On Thu, 7 Sept 2023 at 10:33, Markus Armbruster <armbru@redhat.com> wrote:
> > > > >
> > > > > Kevin Wolf <kwolf@redhat.com> writes:
> > > > >
> > > > > > Am 04.09.2023 um 18:25 hat Daniel P. Berrangé geschrieben:
> > > > > >> By the time of the 8.2.0 release, it will have been 2 years and 6
> > > > > >> releases since we accidentally broke setting of array properties
> > > > > >> for user creatable devices:
> > > > > >>
> > > > > >>   https://gitlab.com/qemu-project/qemu/-/issues/1090
> > > > > >
> > > > > > Oh, nice!
> > > > >
> > > > > Nice?  *Awesome*!
> > > > >
> > > > > > Well, maybe that sounds a bit wrong, but the syntax that was broken was
> > > > > > problematic and more of a hack,
> > > > >
> > > > > A monstrosity, in my opinion.  I tried to strangle it in the crib, but
> > > > > its guardians wouldn't let me.  Can dig up references for the morbidly
> > > > > curious.
> > > > 
> > > > I don't care about the syntax on the command line much (AFAIK that's
> > > > just the rocker device). But the actual feature is used more widely
> > > > within QEMU itself for devices created in C code, which is what it
> > > > was intended for. If you want to get rid of it you need to provide
> > > > an adequate replacement.
> > > 
> > > I have a patch to use QList (i.e. JSON lists) that seems to work for the
> > > rocker case. Now I need to find and update all of those internal
> > > callers. Should grepping for '"len-' find all instances that need to be
> > > changed or are you aware of other ways to access the feature?
> > 
> > IMHO we can just leave the internal only code callers unchanged. I was
> > about to send this patch to prevent usage leaking into user creatable
> > devices:
> > [...]
> 
> The bug report is about a user creatable device (rocker) that doesn't
> work any more, so forbidding the case entirely doesn't really improve
> the situation.

Oh, sure we still need to patch rocker.c to make it use a QAPI friendly
way of accepting the list of devices.

The forbidding patch was just to proactively prevent the bad design
practice recurring in the future.

With regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|