Re: [Qemu-stable] [Qemu-devel] [PATCH v3 2/2] monitor: fix object_del for command-line-created objects

[Michael Roth]

Quoting Markus Armbruster (2016-12-12 09:53:20)
> Michael Roth <address@hidden> writes:
> 
> > Currently objects specified on the command-line are only partially
> > cleaned up when 'object_del' is issued in either HMP or QMP: the
> > object itself is fully finalized, but the QemuOpts are not removed.
> > This results in the following behavior:
> >
> >   x86_64-softmmu/qemu-system-x86_64 -monitor stdio \
> >     -object memory-backend-ram,id=ram1,size=256M
> >
> >   QEMU 2.7.91 monitor - type 'help' for more information
> >   (qemu) object_del ram1
> >   (qemu) object_del ram1
> >   object 'ram1' not found
> >   (qemu) object_add memory-backend-ram,id=ram1,size=256M
> >   Duplicate ID 'ram1' for object
> >   Try "help object_add" for more information
> >
> > which can be an issue for use-cases like memory hotplug.
> >
> > This happens on the HMP side because hmp_object_add() attempts to
> > create a temporary QemuOpts entry with ID 'ram1', which ends up
> > conflicting with the command-line-created entry, since it was never
> > cleaned up during the previous hmp_object_del() call.
> >
> > We address this by adding a check in user_creatable_del(), which
> > is called by both qmp_object_del() and hmp_object_del() to handle
> > the actual object cleanup, to determine whether an option group entry
> > matching the object's ID is present and removing it if it is.
> >
> > Note that qmp_object_add() never attempts to create a temporary
> > QemuOpts entry, so it does not encounter the duplicate ID error,
> > which is why this isn't generally visible in libvirt.
> >
> > Cc: "Dr. David Alan Gilbert" <address@hidden>
> > Cc: Markus Armbruster <address@hidden>
> > Cc: Eric Blake <address@hidden>
> > Cc: Daniel Berrange <address@hidden>
> > Cc: address@hidden
> > Signed-off-by: Michael Roth <address@hidden>
> > Reviewed-by: Daniel P. Berrange <address@hidden>
> > ---
> >  qom/object_interfaces.c    | 7 +++++++
> >  tests/check-qom-proplist.c | 4 ++--
> >  2 files changed, 9 insertions(+), 2 deletions(-)
> >
> > diff --git a/qom/object_interfaces.c b/qom/object_interfaces.c
> > index ded4d84..b462a74 100644
> > --- a/qom/object_interfaces.c
> > +++ b/qom/object_interfaces.c
> > @@ -5,6 +5,7 @@
> >  #include "qapi-visit.h"
> >  #include "qapi/qobject-output-visitor.h"
> >  #include "qapi/opts-visitor.h"
> > +#include "qemu/config-file.h"
> >  
> >  void user_creatable_complete(Object *obj, Error **errp)
> >  {
> > @@ -209,6 +210,12 @@ void user_creatable_del(const char *id, Error **errp)
> >          error_setg(errp, "object '%s' is in use, can not be deleted", id);
> >          return;
> >      }
> > +
> > +    /* if object was defined on the command-line, remove its corresponding
> > +     * option group entry
> > +     */
> 
> Consider adding wings to both ends of multi-line comments.
> 
> > +    qemu_opts_del(qemu_opts_find(qemu_find_opts_err("object", NULL), id));
> 
> qemu_find_opts_err() can't fail.  If it did, qemu_opts_find() would
> crash.  Okay.
> 
> Elsewhere, we use qemu_opts_find(qemu_find_opts(NAME)), which would
> report an error to stderr / monitor before crashing.  Also okay.
> 
> Perhaps the cleanest (but not the tersest) way to skin this cat:
> qemu_opts_find(qemu_find_opts_err(&error_abort)).
> 
> Your choice.
> 
> > +
> >      object_unparent(obj);
> >  }
> >  
> > diff --git a/tests/check-qom-proplist.c b/tests/check-qom-proplist.c
> > index e3f56ca..c5d4b60 100644
> > --- a/tests/check-qom-proplist.c
> > +++ b/tests/check-qom-proplist.c
> > @@ -437,9 +437,9 @@ static void test_dummy_createcmdl(void)
> >       * check for this in user_creatable_del() and remove the QemuOpts if
> >       * it is present.
> >       *
> > -     * FIXME: add an assert to verify that the QemuOpts is cleaned up
> > -     * once the corresponding cleanup code is added.
> > +     * The below check ensures this works as expected.
> >       */
> > +    g_assert(qemu_opts_find(&qemu_object_opts, "dev0") == NULL);
> 
> There's also g_assert_null(), which could result in a more helpful
> message when the assertion fails.  Your choice.

Thanks for the suggestions; I've incorporated all of them into v4.

> 
> >  }
> >  
> >  static void test_dummy_badenum(void)
>