Shan, Haitao
2008-Sep-09 08:58 UTC
[Xen-devel] [PATCH 0/4] CPU online/offline support in Xen
Hi, Keir, The following patches enable Xen to online/offline a physical CPU. This feature will be needed by possibly CPU hot-plug and RAS features. Some code of cpu online/offline is already in Xen, ever since S3 feature was added to Xen. This patch series only adds the missing part. Note: We only add the code to implement the function that actually online/offline a CPU. The patches for tests are only served as a test purpose. We do the tests by hooking online/offline decision to vcpu online/offline path in Dom0. Thus, writes to /sys/devices/system/cpu/cpuX/online will online/offline the vcpu, as well as physical cpu. Here is the description of the patches. cpu_offline.patch Implements cpu offline feature cpu_online.patch Implements cpu online feature cpu_online_offline_test.patch Used for tests -- Xen part cpu_online_offline_kernel.patch Used for tests -- Dom0 kernel part Best Regards Haitao Shan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Christoph Egger
2008-Sep-09 09:06 UTC
Re: [Xen-devel] [PATCH 0/4] CPU online/offline support in Xen
SUN also has code for physical cpu hotplugging. So this is duplicate work now. Please get into contact with each other and merge your patches. Christoph On Tuesday 09 September 2008 10:58:41 Shan, Haitao wrote:> Hi, Keir, > > The following patches enable Xen to online/offline a physical CPU. This > feature will be needed by possibly CPU hot-plug and RAS features. > Some code of cpu online/offline is already in Xen, ever since S3 feature > was added to Xen. This patch series only adds the missing part. > > Note: We only add the code to implement the function that actually > online/offline a CPU. The patches for tests are only served as a test > purpose. We do the tests by hooking online/offline decision to vcpu > online/offline path in Dom0. Thus, writes to > /sys/devices/system/cpu/cpuX/online will online/offline the vcpu, as > well as physical cpu. > > > Here is the description of the patches. > > cpu_offline.patch Implements cpu offline feature > cpu_online.patch Implements cpu online feature > > cpu_online_offline_test.patch Used for tests -- Xen part > cpu_online_offline_kernel.patch Used for tests -- Dom0 kernel part > > > Best Regards > Haitao Shan-- AMD Saxony, Dresden, Germany Operating System Research Center Legal Information: AMD Saxony Limited Liability Company & Co. KG Sitz (Geschäftsanschrift): Wilschdorfer Landstr. 101, 01109 Dresden, Deutschland Registergericht Dresden: HRA 4896 vertretungsberechtigter Komplementär: AMD Saxony LLC (Sitz Wilmington, Delaware, USA) Geschäftsführer der AMD Saxony LLC: Dr. Hans-R. Deppe, Thomas McCoy _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Shan, Haitao
2008-Sep-09 09:08 UTC
RE: [Xen-devel] [PATCH 0/4] CPU online/offline support in Xen
Sorry, I have not seen any patches yet. Are they sent to maillists before? Haitao Shan -----Original Message----- From: Christoph Egger [mailto:Christoph.Egger@amd.com] Sent: 2008年9月9日 17:07 To: xen-devel@lists.xensource.com Cc: Shan, Haitao; Keir Fraser; Gavin Maltby Subject: Re: [Xen-devel] [PATCH 0/4] CPU online/offline support in Xen SUN also has code for physical cpu hotplugging. So this is duplicate work now. Please get into contact with each other and merge your patches. Christoph On Tuesday 09 September 2008 10:58:41 Shan, Haitao wrote:> Hi, Keir, > > The following patches enable Xen to online/offline a physical CPU. This > feature will be needed by possibly CPU hot-plug and RAS features. > Some code of cpu online/offline is already in Xen, ever since S3 feature > was added to Xen. This patch series only adds the missing part. > > Note: We only add the code to implement the function that actually > online/offline a CPU. The patches for tests are only served as a test > purpose. We do the tests by hooking online/offline decision to vcpu > online/offline path in Dom0. Thus, writes to > /sys/devices/system/cpu/cpuX/online will online/offline the vcpu, as > well as physical cpu. > > > Here is the description of the patches. > > cpu_offline.patch Implements cpu offline feature > cpu_online.patch Implements cpu online feature > > cpu_online_offline_test.patch Used for tests -- Xen part > cpu_online_offline_kernel.patch Used for tests -- Dom0 kernel part > > > Best Regards > Haitao Shan-- AMD Saxony, Dresden, Germany Operating System Research Center Legal Information: AMD Saxony Limited Liability Company & Co. KG Sitz (Geschäftsanschrift): Wilschdorfer Landstr. 101, 01109 Dresden, Deutschland Registergericht Dresden: HRA 4896 vertretungsberechtigter Komplementär: AMD Saxony LLC (Sitz Wilmington, Delaware, USA) Geschäftsführer der AMD Saxony LLC: Dr. Hans-R. Deppe, Thomas McCoy _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2008-Sep-09 10:20 UTC
Re: [Xen-devel] [PATCH 0/4] CPU online/offline support in Xen
You can hardly expect to do work in private and then ''surface the submarine'' when others post patches and expect them to do the work to merge with you. If Intel''s patches look acceptable then they will be checked in and Sun will have to do the merge work to get any extra enhancements or features in their patchset to work on top. That is unless you can see strong reason why Sun''s patches are technically better. -- Keir On 9/9/08 10:06, "Christoph Egger" <Christoph.Egger@amd.com> wrote:> > SUN also has code for physical cpu hotplugging. > So this is duplicate work now. > Please get into contact with each other and merge your patches. > > Christoph > > > On Tuesday 09 September 2008 10:58:41 Shan, Haitao wrote: >> Hi, Keir, >> >> The following patches enable Xen to online/offline a physical CPU. This >> feature will be needed by possibly CPU hot-plug and RAS features. >> Some code of cpu online/offline is already in Xen, ever since S3 feature >> was added to Xen. This patch series only adds the missing part. >> >> Note: We only add the code to implement the function that actually >> online/offline a CPU. The patches for tests are only served as a test >> purpose. We do the tests by hooking online/offline decision to vcpu >> online/offline path in Dom0. Thus, writes to >> /sys/devices/system/cpu/cpuX/online will online/offline the vcpu, as >> well as physical cpu. >> >> >> Here is the description of the patches. >> >> cpu_offline.patch Implements cpu offline feature >> cpu_online.patch Implements cpu online feature >> >> cpu_online_offline_test.patch Used for tests -- Xen part >> cpu_online_offline_kernel.patch Used for tests -- Dom0 kernel part >> >> >> Best Regards >> Haitao Shan > >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Gavin Maltby
2008-Sep-09 10:35 UTC
Re: [Xen-devel] [PATCH 0/4] CPU online/offline support in Xen
Hi, Keir Fraser wrote: > You can hardly expect to do work in private and then ''surface the submarine'' > when others post patches and expect them to do the work to merge with you. > If Intel''s patches look acceptable then they will be checked in and Sun will > have to do the merge work to get any extra enhancements or features in their > patchset to work on top. That is unless you can see strong reason why Sun''s > patches are technically better. > Agreed. Sun and Intel have exchanged a few loose emails on our intentions in this area, but no close coordination. We (Sun) aren''t expecting to try to contribute ours as patches to xen-unstable straight-away as we have a broader project to float first, and we also have to take our 3.1.4-based work up to the xen-unstable base. So by all means take in the Intel patch - we have no objections (although we might feel differently when next our base is updated to a later Xen release!). I''ve asked the author of the Sun work to have a look at this proposed patch to see how it fits with our work. Thanks Gavin> -- Keir > > On 9/9/08 10:06, "Christoph Egger" <Christoph.Egger@amd.com> wrote: > >> SUN also has code for physical cpu hotplugging. >> So this is duplicate work now. >> Please get into contact with each other and merge your patches. >> >> Christoph >> >> >> On Tuesday 09 September 2008 10:58:41 Shan, Haitao wrote: >>> Hi, Keir, >>> >>> The following patches enable Xen to online/offline a physical CPU. This >>> feature will be needed by possibly CPU hot-plug and RAS features. >>> Some code of cpu online/offline is already in Xen, ever since S3 feature >>> was added to Xen. This patch series only adds the missing part. >>> >>> Note: We only add the code to implement the function that actually >>> online/offline a CPU. The patches for tests are only served as a test >>> purpose. We do the tests by hooking online/offline decision to vcpu >>> online/offline path in Dom0. Thus, writes to >>> /sys/devices/system/cpu/cpuX/online will online/offline the vcpu, as >>> well as physical cpu. >>> >>> >>> Here is the description of the patches. >>> >>> cpu_offline.patch Implements cpu offline feature >>> cpu_online.patch Implements cpu online feature >>> >>> cpu_online_offline_test.patch Used for tests -- Xen part >>> cpu_online_offline_kernel.patch Used for tests -- Dom0 kernel part >>> >>> >>> Best Regards >>> Haitao Shan >> > >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2008-Sep-09 10:48 UTC
Re: [Xen-devel] [PATCH 0/4] CPU online/offline support in Xen
On 9/9/08 11:35, "Gavin Maltby" <Gavin.Maltby@Sun.COM> wrote:> Keir Fraser wrote: >> You can hardly expect to do work in private and then ''surface the submarine'' >> when others post patches and expect them to do the work to merge with you. >> If Intel''s patches look acceptable then they will be checked in and Sun will >> have to do the merge work to get any extra enhancements or features in their >> patchset to work on top. That is unless you can see strong reason why Sun''s >> patches are technically better. > > Agreed. Sun and Intel have exchanged a few loose emails on our intentions > in this area, but no close coordination. We (Sun) aren''t expecting > to try to contribute ours as patches to xen-unstable straight-away > as we have a broader project to float first, and we also have to > take our 3.1.4-based work up to the xen-unstable base. So by > all means take in the Intel patch - we have no objections > (although we might feel differently when next our base is > updated to a later Xen release!). I''ve asked the author of the > Sun work to have a look at this proposed patch to see how it > fits with our work.Sounds good to me! Cheers, Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Frank van der Linden
2008-Sep-09 19:20 UTC
Re: [Xen-devel] [PATCH 0/4] CPU online/offline support in Xen
Keir Fraser wrote:> You can hardly expect to do work in private and then ''surface the submarine'' > when others post patches and expect them to do the work to merge with you. > If Intel''s patches look acceptable then they will be checked in and Sun will > have to do the merge work to get any extra enhancements or features in their > patchset to work on top. That is unless you can see strong reason why Sun''s > patches are technically better. >Well, not quite in private, you may remember that I did a presentation at the Xen Summit. Anyway, of course there is no point in not accepting or delaying a patch for something that is not available yet. I''ll comment on the submitted patch, I think I see some issues that we found that aren''t addressed in it. - Frank _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2008-Sep-09 19:31 UTC
Re: [Xen-devel] [PATCH 0/4] CPU online/offline support in Xen
On 9/9/08 20:20, "Frank van der Linden" <Frank.Vanderlinden@Sun.COM> wrote:> Keir Fraser wrote: >> You can hardly expect to do work in private and then ''surface the submarine'' >> when others post patches and expect them to do the work to merge with you. >> If Intel''s patches look acceptable then they will be checked in and Sun will >> have to do the merge work to get any extra enhancements or features in their >> patchset to work on top. That is unless you can see strong reason why Sun''s >> patches are technically better. >> > Well, not quite in private, you may remember that I did a presentation > at the Xen Summit.Ah, I had forgotten about that! The machine-check stuff is what stuck in my mind. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel