From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:37765) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fY9Ee-0004eZ-7o for qemu-devel@nongnu.org; Wed, 27 Jun 2018 08:07:45 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fY9EZ-00084B-8S for qemu-devel@nongnu.org; Wed, 27 Jun 2018 08:07:44 -0400 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:33168 helo=mx1.redhat.com) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1fY9EZ-000830-1V for qemu-devel@nongnu.org; Wed, 27 Jun 2018 08:07:39 -0400 Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.rdu2.redhat.com [10.11.54.6]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id B6D97406EA4F for ; Wed, 27 Jun 2018 12:07:38 +0000 (UTC) Date: Wed, 27 Jun 2018 20:07:33 +0800 From: Peter Xu Message-ID: <20180627120733.GD2516@xz-mi> References: <20180620073223.31964-1-peterx@redhat.com> <871sctea4y.fsf@dusky.pond.sub.org> <87tvpoadcc.fsf_-_@dusky.pond.sub.org> <87wouk8vul.fsf@dusky.pond.sub.org> <20180627102043.GD30628@redhat.com> <8760248odg.fsf@dusky.pond.sub.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <8760248odg.fsf@dusky.pond.sub.org> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] monitor: enable OOB by default List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Markus Armbruster Cc: Daniel =?utf-8?B?UC4gQmVycmFuZ8Op?= , qemu-devel@nongnu.org On Wed, Jun 27, 2018 at 01:23:07PM +0200, Markus Armbruster wrote: > Daniel P. Berrang=C3=A9 writes: >=20 > > On Wed, Jun 27, 2018 at 10:41:38AM +0200, Markus Armbruster wrote: > >> Markus Armbruster writes: > >>=20 > >> > Markus Armbruster writes: > >> > > >> >> I fooled around a bit, and I think there are a few lose ends. > >> > [...] > >> >> Talking to a QMP monitor that supports OOB: > >> >> > >> >> $ socat UNIX:test-qmp READLINE,history=3D$HOME/.qmp_history,p= rompt=3D'QMP> ' > >> >> {"QMP": {"version": {"qemu": {"micro": 50, "minor": 12, "majo= r": 2}, "package": "v2.12.0-1703-gb909799463"}, "capabilities": ["oob"]}} > >> >> QMP> { "execute": "qmp_capabilities", "arguments": { "oob": t= rue } } > >> >> {"error": {"class": "GenericError", "desc": "Parameter 'oob' = is unexpected"}} > >> >> QMP> { "execute": "qmp_capabilities", "arguments": { "enable"= : ["oob"] } } > >> >> {"return": {}} > >> >> QMP> { "execute": "query-qmp-schema" } > >> >> {"error": {"class": "GenericError", "desc": "Out-Of-Band capa= bility requires that every command contains an 'id' field"}} > >> >> > >> >> Why does every command require 'id'? > >> > > >> > I found one reason: event COMMAND_DROPPED wants it. Any other rea= son? > >> > > >> > [...] > >>=20 > >> Apropos COMMAND_DROPPED: we send an event rather than an error respo= nse > >> because we may send it out-of-order. Makes sense. > >>=20 > >> However, broadcasting it to all monitors doesn't make sense. We cou= ld > >> use a way to send an event to just one monitor. True. (Sorry for the late responses; I was on Linuxcon China in the past few days) > > > > Worse than that - broadcasting to all monitors is categorically broke= n. > > Different monitors make use the same "id" formatting scheme, so if yo= u > > broadcast COMMAND_DROPPED to a different monitor you might have clash= ing > > "id" and thus incorrectly tell a client its command was dropped when = in > > fact it was processed. You'd have to be fairly unlucky in timing, but > > it could happen. >=20 > Right. Must fix bug. Even more true. >=20 > I'm glad I went over this one more time, and in public! I had a glance at current qmp-spec, it seems that we don't have any restriction currently on "we must send events to all the monitors". Does it mean that we should be free to have per-monitor events starting from this event? My current plan is that I can touch up scripts/qapi/events.py and related stuff to allow QMPEventFuncEmit to take a monitor parameter, then we pass in NULL when we want to send the event to all monitors. Would that work? Thanks, --=20 Peter Xu