Description of problem: Two upstream xen/ia64 patches were inadvertently left behind in RHEL5.x: http://xenbits.xensource.com/xen-unstable.hg?rev/204046d99562 -- adds missing locking -- initialize missing ticks for steal time accounting on resume http://xenbits.xensource.com/xen-unstable.hg?rev/1a010d9444ba -- avoid soft lockup when domain is resumed Both of these should be fixed in RHEL, so I've ported the patches, tested them and will attach them to this bug. Version-Release number of selected component (if applicable): kernel-2.6.18-91.el5.ia64.rpm (and all prior RHEL5 xen/ia64 kernels)
Created attachment 304706 [details] backport of xen-unstable.hg 204046d99562 and 1a010d9444ba This patch is simpler than the original changesets. Changeset 204046d99562 includes a large block of code that runs optionally if xen_ia64_settimefoday_after_resume is passed on the kernel command-line. That's superfluous to the fix so omitted from the port. The same changeset includes a large chunk of code that's commented out with #if 0, also omitted from the port.
Created attachment 329076 [details] ia64 fix time resume Your alternate patch may be simpler, but we like to stay close to upstream where possible. That just makes it easier for us to port additional fixes in the future. I've attached a new patch which is basically the two upstream c/s combined; this should be a pretty faithful backport. Can you give this a once over to make sure it is sane? Thanks, Chris Lalancette
I've uploaded a test kernel that contains this fix (along with several others) to this location: http://people.redhat.com/clalance/virttest Could the original reporter try out the test kernels there, and report back if it fixes the problem? Thanks, Chris Lalancette
Actually, while going back through BZ's, I've found that this is basically a dup of 288511. Closing it as such. Chris Lalancette *** This bug has been marked as a duplicate of bug 288511 ***