From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:36571) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZciEX-0004N4-BW for qemu-devel@nongnu.org; Thu, 17 Sep 2015 19:04:54 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ZciET-0003Rw-8F for qemu-devel@nongnu.org; Thu, 17 Sep 2015 19:04:53 -0400 Received: from e18.ny.us.ibm.com ([129.33.205.208]:57514) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZciET-0003Rm-3d for qemu-devel@nongnu.org; Thu, 17 Sep 2015 19:04:49 -0400 Received: from /spool/local by e18.ny.us.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Thu, 17 Sep 2015 19:04:48 -0400 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable From: Michael Roth In-Reply-To: <55FAE204.8000403@redhat.com> References: <1442194913-26545-1-git-send-email-david@gibson.dropbear.id.au> <1442194913-26545-2-git-send-email-david@gibson.dropbear.id.au> <55F683CB.6000508@ozlabs.ru> <20150914114503.GP2547@voom.fritz.box> <55F6B989.8010900@redhat.com> <55F6D479.4010301@ozlabs.ru> <55F6D897.1090100@redhat.com> <20150916031635.GY2547@voom.fritz.box> <20150917155053.27212.55263@loki> <55FAE204.8000403@redhat.com> Message-ID: <20150917230349.20869.11698@loki> Date: Thu, 17 Sep 2015 18:03:49 -0500 Subject: Re: [Qemu-devel] [RFCv2 1/2] spapr: Remove unnecessary owner field from sPAPRDRConnector List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Paolo Bonzini , David Gibson Cc: Alexey Kardashevskiy , qemu-devel@nongnu.org, qemu-ppc@nongnu.org, agraf@suse.de, bharata@linux.vnet.ibm.com Quoting Paolo Bonzini (2015-09-17 10:53:40) > = > = > On 17/09/2015 17:50, Michael Roth wrote: > > We still need globals for RTAS lookups. I think QOM is our most > > mature/well-tested interface for managing inter-device > > relationships/lookups, but I can understand if using > > root_container/link seems off. To me it seem like a nice > > "freebie" we get from using QOM, and gave us nice guarantees like > > globally unique paths to correspond to globally unique DRC indexes. > = > I think it's okay; but I don't really like looking at ->properties > directly, without an API (which doesn't exist indeed). Yah, rather than copying qmp_qom_list() I think a proper interface would've been warranted. Perhaps an object_link_foreach() that mirrors object_child_foreach()? For maybe just a more generic object_foreach() with a type mask? Another approach would be to re-use or genericize qmp_qom_list() but that seems like extra allocations/cleanup that aren't really necessary outside of QMP/QAPI return values. > = > Paolo >=20