Keir/Jean, I have one question for changeset 17524 "Guest CPUID configuration" (The URL is http://xenbits.xensource.com/staging/xen-unstable.hg?rev/18727843db60). This changeset do CPUID caculation in libxc. My question is, why don''t we have such support for PV guest? Currently PV guest''s CPUID virtualization is hardcoded in emulate_forced_invalid_op(), is it possible to do things like hvm domain? Thanks Yunhong Jiang _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
There''s no reason not to make PV CPUID similarly configurable. -- Keir On 28/5/08 08:19, "Jiang, Yunhong" <yunhong.jiang@intel.com> wrote:> Keir/Jean, I have one question for changeset 17524 "Guest CPUID > configuration" (The URL is > http://xenbits.xensource.com/staging/xen-unstable.hg?rev/18727843db60). > This changeset do CPUID caculation in libxc. My question is, why don''t > we have such support for PV guest? Currently PV guest''s CPUID > virtualization is hardcoded in emulate_forced_invalid_op(), is it > possible to do things like hvm domain? > > Thanks > Yunhong Jiang_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
So how about dom0? Does that mean we need different handler for dom0/PV guest in emulate_forced_invalid_op()? -- Yunhong Jiang Keir Fraser <mailto:keir.fraser@eu.citrix.com> wrote:> There''s no reason not to make PV CPUID similarly configurable. > > -- Keir > > On 28/5/08 08:19, "Jiang, Yunhong" <yunhong.jiang@intel.com> wrote: > >> Keir/Jean, I have one question for changeset 17524 "Guest CPUID >> configuration" (The URL is >> >http://xenbits.xensource.com/staging/xen-unstable.hg?rev/18727843db60).>> This changeset do CPUID caculation in libxc. My question is, whydon''t>> we have such support for PV guest? Currently PV guest''s CPUID >> virtualization is hardcoded in emulate_forced_invalid_op(), is it >> possible to do things like hvm domain? >> >> Thanks >> Yunhong Jiang_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
We''d have to keep a default policy at least for dom0 hardcoded in Xen. -- Xen On 28/5/08 08:52, "Jiang, Yunhong" <yunhong.jiang@intel.com> wrote:> So how about dom0? Does that mean we need different handler for dom0/PV > guest in emulate_forced_invalid_op()? > > -- Yunhong Jiang > > > Keir Fraser <mailto:keir.fraser@eu.citrix.com> wrote: >> There''s no reason not to make PV CPUID similarly configurable. >> >> -- Keir >> >> On 28/5/08 08:19, "Jiang, Yunhong" <yunhong.jiang@intel.com> wrote: >> >>> Keir/Jean, I have one question for changeset 17524 "Guest CPUID >>> configuration" (The URL is >>> >> > http://xenbits.xensource.com/staging/xen-unstable.hg?rev/18727843db60). >>> This changeset do CPUID caculation in libxc. My question is, why > don''t >>> we have such support for PV guest? Currently PV guest''s CPUID >>> virtualization is hardcoded in emulate_forced_invalid_op(), is it >>> possible to do things like hvm domain? >>> >>> Thanks >>> Yunhong Jiang_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Got it. Thanks for your clarification. -- Yunhong Jiang Keir Fraser <mailto:keir.fraser@eu.citrix.com> wrote:> We''d have to keep a default policy at least for dom0 hardcoded in Xen. > > -- Xen > > On 28/5/08 08:52, "Jiang, Yunhong" <yunhong.jiang@intel.com> wrote: > >> So how about dom0? Does that mean we need different handler fordom0/PV>> guest in emulate_forced_invalid_op()? >> >> -- Yunhong Jiang >> >> >> Keir Fraser <mailto:keir.fraser@eu.citrix.com> wrote: >>> There''s no reason not to make PV CPUID similarly configurable. >>> >>> -- Keir >>> >>> On 28/5/08 08:19, "Jiang, Yunhong" <yunhong.jiang@intel.com> wrote: >>> >>>> Keir/Jean, I have one question for changeset 17524 "Guest CPUID >>>> configuration" (The URL is >>>> >>> >> >http://xenbits.xensource.com/staging/xen-unstable.hg?rev/18727843db60).>>>> This changeset do CPUID caculation in libxc. My question is, whydon''t>>>> we have such support for PV guest? Currently PV guest''s CPUID >>>> virtualization is hardcoded in emulate_forced_invalid_op(), is it >>>> possible to do things like hvm domain? >>>> >>>> Thanks >>>> Yunhong Jiang_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel