From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:53836) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gNhjg-0008K0-GF for qemu-devel@nongnu.org; Fri, 16 Nov 2018 12:16:53 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1gNhjf-0007ES-Jj for qemu-devel@nongnu.org; Fri, 16 Nov 2018 12:16:52 -0500 From: Alberto Garcia In-Reply-To: <1d65ed3f-7353-002f-9000-3b824d5892ce@redhat.com> References: <79c52867-2c10-401b-95d9-2d2edd8afa5e@redhat.com> <462138b3-e9e7-29e5-da55-d0ebd626aee7@redhat.com> <2e1b90ae-1a0c-711a-6ef8-3c814335f696@redhat.com> <20181116151834.GA5066@localhost.localdomain> <001ec125-35f3-e1fb-56a7-1ed4a1974cc6@redhat.com> <20181116155106.GD5066@localhost.localdomain> <1d65ed3f-7353-002f-9000-3b824d5892ce@redhat.com> Date: Fri, 16 Nov 2018 18:16:27 +0100 Message-ID: MIME-Version: 1.0 Content-Type: text/plain Subject: Re: [Qemu-devel] KVM Forum block no[td]es List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Max Reitz , Kevin Wolf Cc: Qemu-block , "qemu-devel@nongnu.org" , Markus Armbruster , "Denis V. Lunev" , Vladimir Sementsov-Ogievskiy (I just wanted to reply quickly to this point, I'll read the rest of the e-mail later) On Fri 16 Nov 2018 05:34:08 PM CET, Max Reitz wrote: > > GRAPH_MOD with the meaning that Berto suggested isn't weird or > > complicated to understand. It's only the wrong tool because it > > blocks more than we want to block. But if we didn't care about that, > > it could be just another permission like any other. That's the way that I understood it. Or, more precisely, I'm basing my current code on those assumptions because it's a simple starting point to have something working with the current permission system. But I'm of course open to trying a different approach. > The meaning Berto has suggested (AFAIU) is never taking the permission > at all but just querying whether it is shared by all current parents > before doing a graph change. That would be a quick way to test it, and I agree that it's unorthodox, but it's useful to explain how I understood the meaning of GRAPH_MOD. But you could for example make BDRVReopenState a parent of the BDS (and replace BDRVReopenState.bs with a BdrvChild). Berto