Red Hat Bugzilla – Bug 455701
cpufreq driver get the wrong processor number in domain by ACPI _PSD table
Last modified: 2008-08-15 15:09:55 EDT
Description of problem:
Hardware: Tylersburg HEDT, Nehalem, BIOS SOX5810J.86A.0977.2008.0624.2121.
RHEL5.2 cpufreq driver get the the processor number in domain equal "0" by
ACPI _PSD table. Actually, ACPI _PSD table show the processor number in domain
equal "8". In other hand, FedoraCore 9 can get the correct processor number
in domain equaling "8". Processor number in domain is important parameter for
cpufreq driver and governor do SW/HW coordination.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Boot RHEL5.2
2. cat /sys/devices/system/cpu/cpu0/cpufreq/affected_cpus.
3. affected_cpus show "0".
I validation on the Tylersburg HEDT, the upstream patche can fix this bug. The
commit ID 152c300d007c70c4a1847dad39ecdaba22e7d457.
Created attachment 312018 [details]
Patch for commit ID 152c300d007c70c4a1847dad39ecdaba22e7d457
I can not access the 455516 bug.
This bug almost seems like a dupe of #428909 since the same patch solves both
Build a brew kernel with the patch and have people test against bug 428909. I
will mark this one as a duplicate depending on the test results.
*** This bug has been marked as a duplicate of bug 428909 ***
Are there any BIOS settings which affect this? I am trying to test on HEDT system with 0926 BIOS version
The BIOS on the system is SOX5810J.86A.1171.2008.0717.0926 so it is the 1171 BIOS.
Disregard BIOS revision I specified in Comment 7