Bug 449346

Adam, which timer mode did you use for the patch to fix your problem?

Rik, in my Caneland platform, I get much high hang precent than 20% to boot a SMP PAE RHEL5u1/2 guest. 32e guest doesn't have this issue. 
I remember both timer_mode 1 and 2 would help booting.

It looks like we will also need the following two changesets from xen-unstable, in addition to the changeset above:

changeset:   18554:1420a6649cfa
user:        Keir Fraser <keir.fraser>
date:        Fri Sep 26 17:09:36 2008 +0100
summary:     hvm: Default timer_mode=1 (do not delay virtual time for missed

changeset:   16764:3f26758bcc02
user:        Keir Fraser <keir.fraser>
date:        Fri Jan 18 22:27:51 2008 +0000
summary:     xend: Handle unspecified timer_mode domain platform parameter.

Other timer changesets we may need for HVM time to run correctly:

changeset:   18729:16eede823854
user:        Keir Fraser <keir.fraser>
date:        Tue Oct 28 10:36:22 2008 +0000
summary:     hvm: Do not mess with APIC timer deadline if in one-shot mode.

changeset:   18695:6f74549ac4c5
user:        Keir Fraser <keir.fraser>
date:        Wed Oct 22 12:08:16 2008 +0100
summary:     x86, hvm: Allow 100us periodic virtual timers

changeset:   18694:71c15dfaa12b
user:        Keir Fraser <keir.fraser>
date:        Wed Oct 22 12:04:32 2008 +0100
summary:     Port HPET device model to vpt timer subsystem

changeset:   17716:6c4cab061af4
user:        Keir Fraser <keir.fraser>
date:        Sat May 24 09:27:03 2008 +0100
summary:     hvm: Build guest timers on monotonic system time.

changeset:   16690:01adaec882d4
user:        Keir Fraser <keir.fraser>
date:        Tue Jan 08 14:31:23 2008 +0000
summary:     hvm: time: Fixes to 'SYNC' (no_missed_ticks_pending) timer handling
.

changeset:   16689:66db23ecd562
user:        Keir Fraser <keir.fraser>
date:        Tue Jan 08 13:57:45 2008 +0000
summary:     hvm: hpet: Fix per-timer enable/disable.

Related to bug 307201 ?

Some more HVM HPET candidates:

changeset:   17017:209512f6d89c
user:        Keir Fraser <keir.fraser>
date:        Mon Feb 11 14:45:29 2008 +0000
summary:     x86 hvm: Allow HPET to be configured as a per-domain config option.

changeset:   16707:51aa2f884f64
user:        Keir Fraser <keir.fraser>
date:        Fri Jan 11 11:01:36 2008 +0000
summary:     hvm: hpet: Tidy up hpet_to_ns_limit calculation.

changeset:   16697:1b2be7cf0b7b
user:        Keir Fraser <keir.fraser>
date:        Wed Jan 09 10:32:13 2008 +0000
summary:     hvm: hpet: Clamp period to sane values to prevent excessive looping
 in

changeset:   16693:9ff64d045e61
user:        Keir Fraser <keir.fraser>
date:        Tue Jan 08 16:20:04 2008 +0000
summary:     hvm: hpet: Fix overflow when converting to nanoseconds.
changeset:   16689:66db23ecd562
user:        Keir Fraser <keir.fraser>
date:        Tue Jan 08 13:57:45 2008 +0000
summary:     hvm: hpet: Fix per-timer enable/disable.

changeset:   16486:c00f31f27de6
user:        Keir Fraser <keir.fraser>
date:        Wed Nov 28 13:13:51 2007 +0000
summary:     hvm: Fix 2 type mismatches in vlapic.h and hpet.c for 32-bit build 
Xen

changeset:   16404:ae6f4c7f15cb
user:        Keir Fraser <keir.fraser>
date:        Wed Nov 21 09:49:09 2007 +0000
summary:     hvm: Do not crash guest if it does an unaligned access to an HPET

Created attachment 331184 [details]
patch 1/3 of the tools bit

I am building a kernel now with a bunch of backported patches, CVS branch private-bz449346-branch, http://brewweb.devel.redhat.com/brew/taskinfo?taskID=1682472

Created attachment 331185 [details]
patch 2/3 of the tools bits

Created attachment 331186 [details]
patch 3/3 of the tools bits

This bug can be caused by a combination of two main factors:
- while doing disk IO, one VCPU of an HVM guest can miss timer ticks
- Xen did not re-deliver those missed timer ticks later on, causing clock skew between VCPUs inside an HVM guest

Both of these issues should be resolved with the backport of the AIO disk handling code and upstream Xen 'no missed-tick accounting' timer code. Please test the test RPMs from http://people.redhat.com/riel/.xenaiotime/ and let us know if those (experimental!) test packages resolve the issue.

*** Bug 490760 has been marked as a duplicate of this bug. ***

*** Bug 499276 has been marked as a duplicate of this bug. ***

in kernel-2.6.18-146.el5
You can download this test kernel from http://people.redhat.com/dzickus/el5

Please do NOT transition this bugzilla state to VERIFIED until our QE team
has sent specific instructions indicating when to do so.  However feel free
to provide a comment indicating that this fix has been verified.

~~ Attention - RHEL 5.4 Beta Released! ~~

RHEL 5.4 Beta has been released! There should be a fix present in the Beta release that addresses this particular request. Please test and report back results here, at your earliest convenience. RHEL 5.4 General Availability release is just around the corner!

If you encounter any issues while testing Beta, please describe the issues you have encountered and set the bug into NEED_INFO. If you encounter new issues, please clone this bug to open a new issue and request it be reviewed for inclusion in RHEL 5.4 or a later update, if it is not of urgent severity.

Please do not flip the bug status to VERIFIED. Only post your verification results, and if available, update Verified field with the appropriate value.

Questions can be posted to this bug or your customer or partner representative.

Try this issue on RHEL5.4 beta. 

If using RHEL5u1 PAE image, still see few times guest booting stopped at "waiting for driver initialization".

But doesn't see any issue with RHEL5u4 PAE image.

*** Bug 461640 has been marked as a duplicate of this bug. ***

*** Bug 307201 has been marked as a duplicate of this bug. ***

An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHSA-2009-1243.html

The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days