Is there a particular reason for doing this unconditionally on both 32- and 64-bits in xenolinux? It would seem to me that this feature should be disabled at most in unprivileged kernels... Thanks, Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 9 Jan 2006, at 16:06, Jan Beulich wrote:> Is there a particular reason for doing this unconditionally on both 32- > and 64-bits in xenolinux? It would seem to me that this feature should > be disabled at most in unprivileged kernels... Thanks, JanHow do you mean? Looking at xen-unstable, we clear CR4_DE at the same point as native linux (i.e., we haven''t changed code relating to that control flag at all). It''s not a debug-feature-enable flag anyway -- more a backwards-compatibility flag for accesses to defunct debug registers dr4 and dr5. It''s pretty unimportant. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
>>> Keir Fraser <Keir.Fraser@cl.cam.ac.uk> 09.01.06 18:13:43 >>> > >On 9 Jan 2006, at 16:06, Jan Beulich wrote: > >> Is there a particular reason for doing this unconditionally on both32->> and 64-bits in xenolinux? It would seem to me that this featureshould>> be disabled at most in unprivileged kernels... Thanks, Jan > >How do you mean? Looking at xen-unstable, we clear CR4_DE at the same>point as native linux (i.e., we haven''t changed code relating to that>control flag at all).I don''t mean the cr4 modification; I rather referred to the feature flag, which gets cleared in setup_arch_post.h''s machine_specific_modify_cpu_capabilities().>It''s not a debug-feature-enable flag anyway -- more a >backwards-compatibility flag for accesses to defunct debug registers >dr4 and dr5. It''s pretty unimportant.Not really. Its important function is to allow I/O port breakpoints. (And I think disallowing access to dr4/dr5 isn''t a bad thing, too.) Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 10 Jan 2006, at 07:46, Jan Beulich wrote:> I don''t mean the cr4 modification; I rather referred to the feature > flag, which gets cleared in setup_arch_post.h''s > machine_specific_modify_cpu_capabilities().Ah, I think Christian disabled it at some point to work around a crash later in boot. I expect there is a deeper underlying problem that ought to be addressed, rather than simply disabling the feature. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel