From mboxrd@z Thu Jan 1 00:00:00 1970 From: Anthony PERARD Subject: Re: [PATCH] Plumb through xen-platform device logging Date: Thu, 11 Jul 2013 14:54:16 +0100 Message-ID: <51DEB908.9020809@citrix.com> References: <1372951983-28428-1-git-send-email-paul.durrant@citrix.com> <9AAE0902D5BC7E449B7C8E4E778ABCD006053C@LONPEX01CL01.citrite.net> <1373543014.5453.179.camel@hastur.hellion.org.uk> <9AAE0902D5BC7E449B7C8E4E778ABCD0060AA7@LONPEX01CL01.citrite.net> <1373547056.5453.184.camel@hastur.hellion.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1373547056.5453.184.camel@hastur.hellion.org.uk> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Ian Campbell Cc: Paul Durrant , Stefano Stabellini , "xen-devel@lists.xen.org" List-Id: xen-devel@lists.xenproject.org On 11/07/13 13:50, Ian Campbell wrote: > On Thu, 2013-07-11 at 13:07 +0100, Paul Durrant wrote: > >>> > > Looking at http://wiki.qemu.org/Features/Tracing is the tracing >>> > > interface really the right way to be logging this particular class of >>> > > information? I'd have thought a simple logfile support in the platform >>> > > device would be a much more natural fit. >>> > > >> > >> > That makes sense to me, but whoever coded up the platform device >> > obviously believed tracing to be the correct way to log. I don't know >> > the history of that decision. > I guess either Anthony or Stefano knows. Do you guys know why we log the > platform device I/O port debug via the trace subsystems? It doesn't seem > like a good fit. It seams that I made this choice long time ago. http://marc.info/?l=xen-devel&m=129864357001108&w=2 But I never try to use it, that was maybe not a great choice. After a quick look into the qemu tree, using the trace thing those not seams bad as well. There is just few step: - compile qemu with traces enable - adding to the vm-xl-config this: device_model_args_hvm = [ '-trace', 'events=/tmp/traces' ] (with "xen_platform_log" in /tmp/traces) So, about the patch, I don't feel a good idea to have it enable all the time for all the guest. Also, QEMU will refuse to start if it's compiled without trace support. > A better fit would be the qemu chr subsystem (I think that's the name, I > mean the thing which lets you direct serial/parallel etc to > file,tcp,sockets etc etc.) This can maybe be done using a property, which could be set via the command line by using the -device options. -- Anthony PERARD