All, Currently, RHEL-4.7 and RHEL-5.3(preview) KVM guests will crash on bootup on Intel machines running F-9 kernels. This is because of a bug in the way that KVM emulates performance counter MSRs. This was fixed upstream by this commit:
http://git2.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h...
However, it fell into a strange window and missed both 2.6.25 and 2.6.26, but is now in 2.6.27 (which means F-10 will be just fine). In the interest of helping out the Fedora users (especially for Bugzilla bug 462177 and duplicates), I backported this very simple patch to the current head of F-9 CVS (2.6.26.5). The patch is attached. If there are no objections, can someone please commit this?
Thanks,
On Wed, Sep 17, 2008 at 04:32:05PM +0200, Chris Lalancette wrote:
All, Currently, RHEL-4.7 and RHEL-5.3(preview) KVM guests will crash on bootup on Intel machines running F-9 kernels. This is because of a bug in the way that KVM emulates performance counter MSRs. This was fixed upstream by this commit:
http://git2.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h...
However, it fell into a strange window and missed both 2.6.25 and 2.6.26, but is now in 2.6.27 (which means F-10 will be just fine). In the interest of helping out the Fedora users (especially for Bugzilla bug 462177 and duplicates), I backported this very simple patch to the current head of F-9 CVS (2.6.26.5). The patch is attached. If there are no objections, can someone please commit this?
Attached where? Has the Zimbra patch goblin eaten it again?
josh
Josh Boyer wrote:
On Wed, Sep 17, 2008 at 04:32:05PM +0200, Chris Lalancette wrote:
All, Currently, RHEL-4.7 and RHEL-5.3(preview) KVM guests will crash on bootup on Intel machines running F-9 kernels. This is because of a bug in the way that KVM emulates performance counter MSRs. This was fixed upstream by this commit:
http://git2.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h...
However, it fell into a strange window and missed both 2.6.25 and 2.6.26, but is now in 2.6.27 (which means F-10 will be just fine). In the interest of helping out the Fedora users (especially for Bugzilla bug 462177 and duplicates), I backported this very simple patch to the current head of F-9 CVS (2.6.26.5). The patch is attached. If there are no objections, can someone please commit this?
Attached where? Has the Zimbra patch goblin eaten it again?
Yes, apparently it has, since my outgoing mail shows it attached. Hm, let's try this another way then; patch is available here:
http://people.redhat.com/clalance/linux-2.6.26-intel-msr-backport.patch
Thanks,
kernel@lists.fedoraproject.org
-
Chris Lalancette -
Josh Boyer -
Kyle McMartin