Just a curious question, does anyone still care xen/UP? If yes, current xen-unstable.hg failed to build quickly after undef CONFIG_SMP since many structures include types defined only when CONFIG_SMP is on. Just realize this issue when cleanup some IA64 code recently, where CONFIG_SMP is still unstable and thus UP only. Is there any benefit to have xen/UP? Reducing image size is the immediate answer in my head... Thanks, Kevin _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 23 Feb 2006, at 08:51, Tian, Kevin wrote:> Just a curious question, does anyone still care xen/UP? If yes, current > xen-unstable.hg > failed to build quickly after undef CONFIG_SMP since many structures > include types > defined only when CONFIG_SMP is on. > > Just realize this issue when cleanup some IA64 code recently, where > CONFIG_SMP is still unstable and thus UP only. > > Is there any benefit to have xen/UP? Reducing image size is the > immediate answer in my head...It''s not supported for xen/x86 at least. Dynamically adding LOCK prefixes if the system turns out to be multiprocessor is the only optimisation I think would be worthwhile. But really multi-processor/core/thread is what we care about. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
>From: Keir Fraser [mailto:Keir.Fraser@cl.cam.ac.uk] >Sent: 2006年2月23日 17:47 > >On 23 Feb 2006, at 08:51, Tian, Kevin wrote: > >> Just a curious question, does anyone still care xen/UP? If yes, current >> xen-unstable.hg >> failed to build quickly after undef CONFIG_SMP since many structures >> include types >> defined only when CONFIG_SMP is on. >> >> Just realize this issue when cleanup some IA64 code recently, where >> CONFIG_SMP is still unstable and thus UP only. >> >> Is there any benefit to have xen/UP? Reducing image size is the >> immediate answer in my head... > >It''s not supported for xen/x86 at least. Dynamically adding LOCK >prefixes if the system turns out to be multiprocessor is the only >optimisation I think would be worthwhile. But really >multi-processor/core/thread is what we care about. > > -- KeirYep, agree. It''s just raised out in case anyone has some special/meaningful usage model coming. :-) Thanks, Kevin _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser wrote:> On 23 Feb 2006, at 08:51, Tian, Kevin wrote: > >> Just a curious question, does anyone still care xen/UP? If yes, >> current xen-unstable.hg failed to build quickly after undef >> CONFIG_SMP since many structures include types defined only when >> CONFIG_SMP is on. >> >> Just realize this issue when cleanup some IA64 code recently, where >> CONFIG_SMP is still unstable and thus UP only. >> >> Is there any benefit to have xen/UP? Reducing image size is the >> immediate answer in my head... > > It''s not supported for xen/x86 at least. Dynamically adding LOCK > prefixes if the system turns out to be multiprocessor is the only > optimisation I think would be worthwhile. But really > multi-processor/core/thread is what we care about. > > -- Keir >That''s right. I think the right thing to do is to set CONFIG_SMP by default forcing people to stabilize IA-64 Xen ASAP, rather than keeping it UP only. Jun --- Intel Open Source Technology Center _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel