I have ported xenoprof to linux-pvops. - All the changes/clean-ups are limited to linux modules (except a xen interface for kernel space mappings). - Calls are switched to xen by checking for xen_initial_domain. Is there a better way of handling such modules with pvops ? - What is the best way to go about checking in this code ? Do I submit all the patches to jeremy''s tree (cause these changes wouldn''t make sense anyway without privileged domain support) ? -dulloor _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 06/29/09 16:10, Dulloor wrote:> I have ported xenoprof to linux-pvops. > > - All the changes/clean-ups are limited to linux modules (except a xen > interface for kernel space mappings). > - Calls are switched to xen by checking for xen_initial_domain. Is > there a better way of handling such modules with pvops ? > > - What is the best way to go about checking in this code ? Do I submit > all the patches to jeremy''s tree (cause these changes wouldn''t make > sense anyway without privileged domain support) ?Yes. The best thing would be if you can set up a git branch somewhere, which I can then pull into my repo. J _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hi Dulloor, Which commit of jeremy's pv-ops tree and changeset of Xen unstable are you using for your porting? pv-ops dom0 cannot boot for me. You can find information in another thread "Xen cannot boot with pvops dom0". Regards, Weidong ________________________________ From: xen-devel-bounces@lists.xensource.com [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Dulloor Sent: 2009年6月30日 7:10 To: xen-devel@lists.xensource.com; Jeremy Fitzhardinge Subject: [Xen-devel] xenoprof on linux pvops I have ported xenoprof to linux-pvops. - All the changes/clean-ups are limited to linux modules (except a xen interface for kernel space mappings). - Calls are switched to xen by checking for xen_initial_domain. Is there a better way of handling such modules with pvops ? - What is the best way to go about checking in this code ? Do I submit all the patches to jeremy's tree (cause these changes wouldn't make sense anyway without privileged domain support) ? -dulloor _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On this machine, I am on a slightly older pv-ops commit. xen changeset is more recent. commit 8ab8a780233482cab6796bc680652cee59be827b Merge: f749557... 0a834a9... Author: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Date: Thu May 21 15:16:53 2009 -0700 changeset: 19806:703ced548925 tag: tip user: Keir Fraser <keir.fraser@citrix.com> date: Fri Jun 19 08:45:55 2009 +0100 summary: VT-d: fix MSI source-id of interrupt remapping -dulloor 2009/6/30 Han, Weidong <weidong.han@intel.com>> Hi Dulloor, > > Which commit of jeremy''s pv-ops tree and changeset of Xen unstable are you > using for your porting? pv-ops dom0 cannot boot for me. You can find > information in another thread "Xen cannot boot with pvops dom0". > > Regards, > Weidong > > > ------------------------------ > *From:* xen-devel-bounces@lists.xensource.com [mailto: > xen-devel-bounces@lists.xensource.com] *On Behalf Of *Dulloor > *Sent:* 2009年6月30日 7:10 > *To:* xen-devel@lists.xensource.com; Jeremy Fitzhardinge > *Subject:* [Xen-devel] xenoprof on linux pvops > > I have ported xenoprof to linux-pvops. > > - All the changes/clean-ups are limited to linux modules (except a xen > interface for kernel space mappings). > - Calls are switched to xen by checking for xen_initial_domain. Is there a > better way of handling such modules with pvops ? > > - What is the best way to go about checking in this code ? Do I submit all > the patches to jeremy''s tree (cause these changes wouldn''t make sense anyway > without privileged domain support) ? > > -dulloor > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel > >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Thanks. I will try your version. Regards, Weidong ________________________________ From: xen-devel-bounces@lists.xensource.com [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Dulloor Sent: 2009年6月30日 14:53 To: Han, Weidong Cc: Jeremy Fitzhardinge; xen-devel@lists.xensource.com Subject: Re: [Xen-devel] xenoprof on linux pvops On this machine, I am on a slightly older pv-ops commit. xen changeset is more recent. commit 8ab8a780233482cab6796bc680652cee59be827b Merge: f749557... 0a834a9... Author: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com<mailto:jeremy.fitzhardinge@citrix.com>> Date: Thu May 21 15:16:53 2009 -0700 changeset: 19806:703ced548925 tag: tip user: Keir Fraser <keir.fraser@citrix.com<mailto:keir.fraser@citrix.com>> date: Fri Jun 19 08:45:55 2009 +0100 summary: VT-d: fix MSI source-id of interrupt remapping -dulloor 2009/6/30 Han, Weidong <weidong.han@intel.com<mailto:weidong.han@intel.com>> Hi Dulloor, Which commit of jeremy's pv-ops tree and changeset of Xen unstable are you using for your porting? pv-ops dom0 cannot boot for me. You can find information in another thread "Xen cannot boot with pvops dom0". Regards, Weidong ________________________________ From: xen-devel-bounces@lists.xensource.com<mailto:xen-devel-bounces@lists.xensource.com> [mailto:xen-devel-bounces@lists.xensource.com<mailto:xen-devel-bounces@lists.xensource.com>] On Behalf Of Dulloor Sent: 2009年6月30日 7:10 To: xen-devel@lists.xensource.com<mailto:xen-devel@lists.xensource.com>; Jeremy Fitzhardinge Subject: [Xen-devel] xenoprof on linux pvops I have ported xenoprof to linux-pvops. - All the changes/clean-ups are limited to linux modules (except a xen interface for kernel space mappings). - Calls are switched to xen by checking for xen_initial_domain. Is there a better way of handling such modules with pvops ? - What is the best way to go about checking in this code ? Do I submit all the patches to jeremy's tree (cause these changes wouldn't make sense anyway without privileged domain support) ? -dulloor _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com<mailto:Xen-devel@lists.xensource.com> http://lists.xensource.com/xen-devel _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
> On 06/29/09 16:10, Dulloor wrote: > > I have ported xenoprof to linux-pvops. > > > > - All the changes/clean-ups are limited to linux modules > (except a xen > > interface for kernel space mappings). > > - Calls are switched to xen by checking for xen_initial_domain. Is > > there a better way of handling such modules with pvops ?On my AMD system, I''m seeing warning from xen_apic_write coming from perf_counters_lapic_init(). Do your patches resolve the issue? -Mark Langsdorf Operating System Research Center AMD _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
I have tested on multiple Intel platforms. I didn''t see this issue. I don''t have an AMD. Jeremy, I am away this week. I will give you access to my git repo by mid next week. -dulloor On Jul 1, 2009, at 9:46 AM, "Langsdorf, Mark" <mark.langsdorf@amd.com> wrote:>> On 06/29/09 16:10, Dulloor wrote: >>> I have ported xenoprof to linux-pvops. >>> >>> - All the changes/clean-ups are limited to linux modules >> (except a xen >>> interface for kernel space mappings). >>> - Calls are switched to xen by checking for xen_initial_domain. Is >>> there a better way of handling such modules with pvops ? > > On my AMD system, I''m seeing warning from xen_apic_write > coming from perf_counters_lapic_init(). Do your patches > resolve the issue? > > -Mark Langsdorf > Operating System Research Center > AMD >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Sorry, I was away for a while. Jeremy, I have forwarded access details to you in another email. -dulloor On Thu, Jul 2, 2009 at 3:38 AM, Dulloor <dulloor@gmail.com> wrote:> I have tested on multiple Intel platforms. I didn''t see this issue. I don''t > have an AMD. > > Jeremy, I am away this week. I will give you access to my git repo by mid > next week. > > -dulloor > > > On Jul 1, 2009, at 9:46 AM, "Langsdorf, Mark" <mark.langsdorf@amd.com> > wrote: > > On 06/29/09 16:10, Dulloor wrote: >>> >>>> I have ported xenoprof to linux-pvops. >>>> >>>> - All the changes/clean-ups are limited to linux modules >>>> >>> (except a xen >>> >>>> interface for kernel space mappings). >>>> - Calls are switched to xen by checking for xen_initial_domain. Is >>>> there a better way of handling such modules with pvops ? >>>> >>> >> On my AMD system, I''m seeing warning from xen_apic_write >> coming from perf_counters_lapic_init(). Do your patches >> resolve the issue? >> >> -Mark Langsdorf >> Operating System Research Center >> AMD >> >>_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel