From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:53859) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1adKqK-0004R0-FB for qemu-devel@nongnu.org; Tue, 08 Mar 2016 11:50:45 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1adKqH-0007qt-4k for qemu-devel@nongnu.org; Tue, 08 Mar 2016 11:50:44 -0500 Received: from mx1.redhat.com ([209.132.183.28]:38878) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1adKqG-0007qj-LS for qemu-devel@nongnu.org; Tue, 08 Mar 2016 11:50:41 -0500 Received: from int-mx10.intmail.prod.int.phx2.redhat.com (int-mx10.intmail.prod.int.phx2.redhat.com [10.5.11.23]) by mx1.redhat.com (Postfix) with ESMTPS id C71F135124C for ; Tue, 8 Mar 2016 16:50:39 +0000 (UTC) Date: Tue, 8 Mar 2016 16:50:35 +0000 From: "Daniel P. Berrange" Message-ID: <20160308165035.GC5010@redhat.com> References: <1457420446-25276-1-git-send-email-peterx@redhat.com> <1457420446-25276-2-git-send-email-peterx@redhat.com> <87pov5e5w3.fsf@blackfin.pond.sub.org> <20160308101944.GD5807@noname.str.redhat.com> <56DEFC16.6010100@redhat.com> <20160308163031.GG5807@noname.str.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20160308163031.GG5807@noname.str.redhat.com> Subject: Re: [Qemu-devel] [PATCH 1/8] qdict: fix unbounded stack for qdict_array_entries Reply-To: "Daniel P. Berrange" List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Kevin Wolf Cc: qemu-devel@nongnu.org, Peter Xu , Markus Armbruster , pbonzini@redhat.com, Luiz Capitulino On Tue, Mar 08, 2016 at 05:30:31PM +0100, Kevin Wolf wrote: > Am 08.03.2016 um 17:21 hat Eric Blake geschrieben: > > On 03/08/2016 03:19 AM, Kevin Wolf wrote: > > > Am 08.03.2016 um 09:22 hat Markus Armbruster geschrieben: > > >> Cc: Kevin, because he added the array in question. > > >> > > >> Peter Xu writes: > > >> > > > > >> Unrelated to your patch: I think we've pushed QDict use father than > > >> sensible. Encoding multiple keys in a string so you can use a flat > > >> associative array as your catch-all data structure is appropriate in > > >> AWK, but in C? Not so much... > > > > > > We'll always to that, because it's the command line syntax. What you may > > > criticise is that we convert QAPI objects to the command line > > > representation instead of the other way around, but changing that (which > > > I think would be far from trivial, for relatively little use) wouldn't > > > get rid of this kind of key parsing, but just move it a bit closer to > > > the command line handling. > > > > I actually WANT us to try that conversion (a great GSoC project, if > > someone wants it) for 2.7. We already did it for SocketAddress, and it > > makes the code easier to maintain when you can just access foo->data > > instead of doing qdict_lookup(foo, "data") all over the place. > > I think you're underestimating the difference in difficulty between > using SocketAddress everywhere and using BlockdevOptions everywhere, > which is the reason why I never even put it on my todo list. > > Of course, I would be fine with your trying anyway, but it would > probably be fairer if we not let a poor student fail with this. I think a more sensible/practical first step would be to change the block layer to using the nested QDicts as its primary representation, instead of the flat QDicts. In APIs which get a flat qdict (CLI QemuOpts & HMP), it should really immediately crumple it into a nested QDict before usage. Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|