Travis Betak
2006-Aug-25 17:28 UTC
[Xen-devel] [PATCH] Lazy save/restore of SVM debug registers
[HVM][SVM] Allow SVM to take advantage of flag_dr_dirty. Signed-off-by: Travis Betak <travis.betak@amd.com> _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
George Dunlap
2006-Aug-27 03:47 UTC
Re: [Xen-devel] [PATCH] Lazy save/restore of SVM debug registers
Hey Travis -- two questions: * This is a safe operation? All debug registers from the guest are turned off before entering the hypervisor? * Is there a reason you don''t save & restore DR6? -George On 8/25/06, Travis Betak <travis.betak@amd.com> wrote:> [HVM][SVM] Allow SVM to take advantage of flag_dr_dirty. > > Signed-off-by: Travis Betak <travis.betak@amd.com> > > _______________________________________________ > 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
Travis Betak
2006-Aug-27 06:11 UTC
Re: [Xen-devel] [PATCH] Lazy save/restore of SVM debug registers
Hi, George, On Sat, 26 Aug 2006, George Dunlap wrote:> Hey Travis -- two questions: > * This is a safe operation? All debug registers from the guest are > turned off before entering the hypervisor?I don''t quite follow although there probably is something that''s not qutie right. Which part are you referring to? After the initial DR write? And as a side note, in SVM, the enable bits in DR7 are cleared on a #VMEXIT.> * Is there a reason you don''t save & restore DR6?In SVM, DR6 is saved along with DR7 in the VMCB on #VMEXITs. --travis _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
George Dunlap
2006-Aug-27 21:55 UTC
Re: [Xen-devel] [PATCH] Lazy save/restore of SVM debug registers
On 8/27/06, Travis Betak <travis.betak@amd.com> wrote:> Hi, George, > > On Sat, 26 Aug 2006, George Dunlap wrote: > > > Hey Travis -- two questions: > > * This is a safe operation? All debug registers from the guest are > > turned off before entering the hypervisor? > > I don''t quite follow although there probably is something that''s not > qutie right. Which part are you referring to? After the initial DR > write? > > And as a side note, in SVM, the enable bits in DR7 are cleared on a > #VMEXIT. > > In SVM, DR6 is saved along with DR7 in the VMCB on #VMEXITs.That''s what I wanted to check. Thanks! -George _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel