Christopher S. Aker
2008-Mar-25 01:35 UTC
[Xen-devel] pv_ops: 2.6.24.3 - no vm86_info: BAD
Xen 3.2, 64bit dom0 = 2.6.16.33 PAE domU = 2.6.24.3 pv_ops EXT3-fs: mounted filesystem with ordered data mode. VFS: Mounted root (ext3 filesystem) readonly. Freeing unused kernel memory: 300k freed init[1] general protection eip:0 esp:0 error:0 no vm86_info: BAD Kernel panic - not syncing: Attempted to kill init! The domU''s distribution is Ubuntu 7.10. It boots Debian 4.0 just fine, so I''m guessing some libc barf. -Chris _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Mon, 2008-03-24 at 21:35 -0400, Christopher S. Aker wrote:> Xen 3.2, 64bit > dom0 = 2.6.16.33 PAE > domU = 2.6.24.3 pv_ops > > EXT3-fs: mounted filesystem with ordered data mode. > VFS: Mounted root (ext3 filesystem) readonly. > Freeing unused kernel memory: 300k freed > init[1] general protection eip:0 esp:0 error:0 > no vm86_info: BAD > Kernel panic - not syncing: Attempted to kill init! > > The domU''s distribution is Ubuntu 7.10. It boots Debian 4.0 just fine, > so I''m guessing some libc barf.If I remember correctly this was due to the generic parts of the kernel trying to use sysenter because the hypervisor claimed to support it in 32on64 mode (by setting SEP in cpuid). However the Xen kernel side code wasn''t present to support it. You can try booting with "nosep" on your command line or grabbing the git changeset d40e705903397445c6861a0a56c23e5b2e8f9b9a (which I thought had been put forward for the stable kernel but I don''t see it there now). Lastly you could also try b8228d191b5ff7e37525846d187807d162e30464 from the x86.git#testing branch which actually implements sysenter support, I don''t know which other changesets that one depends on though. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2008-Mar-25 15:04 UTC
Re: [Xen-devel] pv_ops: 2.6.24.3 - no vm86_info: BAD
Ian Campbell wrote:> If I remember correctly this was due to the generic parts of the kernel > trying to use sysenter because the hypervisor claimed to support it in > 32on64 mode (by setting SEP in cpuid). However the Xen kernel side code > wasn''t present to support it. >Yep, that''s right.> You can try booting with "nosep" on your command line or grabbing the > git changeset d40e705903397445c6861a0a56c23e5b2e8f9b9a (which I thought > had been put forward for the stable kernel but I don''t see it there > now). >Yes, I should make sure that gets pushed out, along with the other bugfix. Booting with "nosep" is a good workaround for the meantime.> Lastly you could also try b8228d191b5ff7e37525846d187807d162e30464 from > the x86.git#testing branch which actually implements sysenter support, I > don''t know which other changesets that one depends on though. >I think its fairly independent of everything else. J _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel