Bug 456437
| Summary: | [RHEL5.2-Z][kernel-xen] powernow identifies the wrong number of processors. | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 5 | Reporter: | Jeff Burke <jburke> |
| Component: | kernel-xen | Assignee: | Miroslav Rezanina <mrezanin> |
| Status: | CLOSED ERRATA | QA Contact: | Martin Jenner <mjenner> |
| Severity: | low | Docs Contact: | |
| Priority: | low | ||
| Version: | 5.2.z | CC: | bhavna.sarathy, bnagendr, duck, dzickus, jpirko, prarit, riel, syeghiay, xen-maint |
| Target Milestone: | rc | Keywords: | Reopened |
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| URL: | http://rhts.redhat.com/testlogs/25754/94808/798424/DMSGDIFF.yK3658 | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2009-09-02 09:00:21 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 492568 | ||
|
Description
Jeff Burke
2008-07-23 16:07:20 UTC
Will re-assign to jdeguara when account is active on BZ. P. RHEL 5.3 has some fixes in this area. Jeff, does the bug happen with the 5.3 kernel-xen? Rik, It does happen with the 5.3 kernel-xen http://rhts.redhat.com/testlogs/40043/136743/1147881/DMSGDIFF.Xr3631.gz =================================================================== diff -u /mnt/testarea/boot.kernel-2.6.18-126.el5 /mnt/testarea/boot.kernel-xen-2.6.18-128.el5 =================================================================== --- /mnt/testarea/boot.kernel-2.6.18-126.el5 2008-12-17 18:12:32.000000000 -0500 +++ /mnt/testarea/boot.kernel-xen-2.6.18-128.el5 2008-12-17 18:17:46.000000000 -0500 @@ -1,31 +1,14 @@ -Linux version 2.6.18-126.el5 (mockbuild.redhat.com) (gcc version 4.1.2 20071124 (Red Hat 4.1.2-42)) #1 SMP Mon Dec 8 18:33:05 EST 2008 +Linux version 2.6.18-128.el5xen (mockbuild.redhat.com) (gcc version 4.1.2 20080704 (Red Hat 4.1.2-44)) #1 SMP Wed Dec 17 12:22:24 EST 2008 -powernow-k8: Found 4 Dual-Core AMD Opteron(tm) Processor 8220 SE processors (8 cpu cores) (version 2.20.00) +powernow-k8: Found 8 Dual-Core AMD Opteron(tm) Processor 8220 SE processors (8 cpu cores) (version 2.20.00) Bhavna, any update on this? Are you working on it? I recall fixing this issue in bare metal with the following snippet (sorry I couldn't find the exact patch on the maling list archive)
#ifdef CONFIG_SMP
printk(KERN_INFO PFX "Found %d %s "
"processors (%d cpu cores) (" VERSION ")\n",
supported_cpus/cpu_data[0].booted_cores,
boot_cpu_data.x86_model_id, supported_cpus);
#else
printk(KERN_INFO PFX "Found 1 %s "
"processors (%d cpu cores) (" VERSION ")\n",
boot_cpu_data.x86_model_id, supported_cpus);
#endif
This identifies the number of processors and not the number of cores. AFAICT, there is no Xen specific powernow-k8 driver, it uses the dom0 driver in arch/i386/kernel/cpu/cpufreq/powernow-k8.c, and the fix that went in RHEL5.1 should apply.
Am I missing something?
Bhavna
Well from the problem description the cpu_data[0].booted_cores would seem to be wrong on the Xen kernel. If it's 2 then the output would be "4 dual core..." and "8 cpu cores". We are getting 8 and 8 on Xen. I have looked through the Xen core enumeration code and noted right away that the code base was taken from Linux i386 and not x86_64. Majority of the code is identical, with no major missing pieces, except for this snippet that is
mostly commented. Bill, can you CC comeone who can comment on this?
/*
* We need to hold call_lock, so there is no inconsistency
* between the time smp_call_function() determines number of
* IPI receipients, and the time when the determination is made
* for which cpus receive the IPI. Holding this
* lock helps us to not include this cpu in a currently in progress
* smp_call_function().
*/
/*lock_ipi_call_lock();*/
cpu_set(smp_processor_id(), cpu_online_map);
/*unlock_ipi_call_lock();*/
/*per_cpu(cpu_state, smp_processor_id()) = CPU_ONLINE;*/
This behavior is caused by Xen functionality. Xen takes true topology and provides to each dom (including dom0) only one-cored proccesors. So dom0 gets in this case 8 processors with 1 core each. However, this processors have original name. Should be behavior of xen smp driver modified to see original number of cores when running in dom0?? Yes it should be fixed because it is incorrect. It is also inconsistent with a standard kernel. Note: Comment #7 in kernel-2.6.18-148.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. 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 |