From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:50045) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1T3o2C-0004P2-23 for qemu-devel@nongnu.org; Tue, 21 Aug 2012 08:58:20 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1T3o27-0000ck-WD for qemu-devel@nongnu.org; Tue, 21 Aug 2012 08:58:15 -0400 Received: from mx1.redhat.com ([209.132.183.28]:1826) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1T3o27-0000cc-Mf for qemu-devel@nongnu.org; Tue, 21 Aug 2012 08:58:11 -0400 Date: Tue, 21 Aug 2012 09:58:52 -0300 From: Luiz Capitulino Message-ID: <20120821095852.29a392c3@doriath.home> In-Reply-To: <20120820193434.GN16157@illuin> References: <1345223489.7008.YahooMailNeo@web122203.mail.ne1.yahoo.com> <20120820105429.5766948f@doriath.home> <20120820193434.GN16157@illuin> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] qemu-ga : Guest Agent : Windows 2008 : Unknown command guest-fsfreeze-freeze List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Michael Roth Cc: desi babu , "qemu-devel@nongnu.org" On Mon, 20 Aug 2012 14:34:34 -0500 Michael Roth wrote: > On Mon, Aug 20, 2012 at 10:54:29AM -0300, Luiz Capitulino wrote: > > On Fri, 17 Aug 2012 10:11:29 -0700 (PDT) > > desi babu wrote: > >=20 > > > Guest-Agent : Windows 2008 Error : Relase 1.1.90 > > >=20 > > > error : internal error unable to execute QEMU command 'guest-fsfreeze= -freeze': this feature or command is not currently supported. > >=20 > > That's correct, fsfreze is not supported on Windows. > >=20 > > > Guest-info shows the command is available. =A0Is there any informatio= n available on the list of commands supported inside Windows ?=A0Appreciate= =A0if you have any pointers. > >=20 > > That's a qemu-ga bug. CC'ing Michael to check if he has a fix in mind f= or this. > >=20 >=20 > I've been thinking about this one for a while. It's considered expected > behavior, but I realize it sucks for discoverability. I doubt we want to > do platform-specific QAPI schema definitions, so the only option I can > think of is some kind of [de-]registration mechanism where we can mark > commands as being not available for a particular build/platform in the > cases where we stub out command implementations. Yes, I guess the first idea I had is similar. I thought about adding PlatformOps and let win32 and posix register the commands they implement. Then, qmp_guest_info() could check if the in-use PlatformOps supports the command before adding it to the supported list. > I think we can expose this to existing clients by no longer listing comma= nds > marked as unsupported in the list provided by the guest-info command. It > should "just work". Can probably do it for 1.3. For now, clients will > have to catch it in the error-handling path. Agreed it's 1.3 material.