Bug 1268604 - Cannot use -p option for over-256 processor systems
Summary: Cannot use -p option for over-256 processor systems
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Fedora EPEL
Classification: Fedora
Component: msr-tools
Version: epel7
Hardware: x86_64
OS: Linux
unspecified
medium
Target Milestone: ---
Assignee: Greg Bailey
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2015-10-03 13:19 UTC by Seiichi Ikarashi
Modified: 2015-11-08 17:52 UTC (History)
1 user (show)

Fixed In Version: msr-tools-1.3-3.el7
Clone Of:
Environment:
Last Closed: 2015-11-08 17:52:02 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)

Description Seiichi Ikarashi 2015-10-03 13:19:06 UTC
Description of problem:
-p option is limited to processor number 255 intentionally but
nowadays, over-256 processor systems are coming.
Let's expand the max limit to, say, 65535 or more.

Version-Release number of selected component (if applicable):
msr-tools-1.3-1.el7

How reproducible:
Always

Steps to Reproduce:
1. Install msr-tools into a system that has 257 processors or more.
2. Use rdmsr command with "-p 256" option.

Actual results:
rdmsr exits with errno 127.

Expected results:
rdmsr should print the MSR value from CPU 256.

Additional info:

A sample fix is
https://github.com/sikarash/msr-tools/commit/3bfa9f712203bd0ad029df8489d2122bd51e0a7b

Comment 1 Greg Bailey 2015-10-03 13:40:37 UTC
Wouldn't we want a similar patch to wrmsr.c and cpuid.c?  It looks like your pull request on github only addresses rdmsr.c.

Comment 2 Seiichi Ikarashi 2015-10-04 05:12:10 UTC
Greg,

(In reply to Greg Bailey from comment #1)
> Wouldn't we want a similar patch to wrmsr.c and cpuid.c?  It looks like your
> pull request on github only addresses rdmsr.c.

You are correct. They are necessary.

https://github.com/sikarash/msr-tools/commit/34f8aa57d4fb571356fcefc5b1b0303af10d0d31
https://github.com/sikarash/msr-tools/commit/a6d856368a354532bf0218dfc334f4e14673271d

Comment 3 Seiichi Ikarashi 2015-10-04 14:21:27 UTC
I was wrong.
The value 65535 is nonsense because the variable is type of (int),
and also the theoretical max logical CPUs of RHEL7 is 5120.
https://access.redhat.com/articles/rhel-limits

So 5120 is enough here.

Comment 4 Seiichi Ikarashi 2015-10-04 14:24:31 UTC
I was wrong again, 2,147,483,647 is the maximum of signed int for x86_64.
Still, 5120 is enough for RHEL7.

Comment 5 Seiichi Ikarashi 2015-10-08 15:57:32 UTC
Greg, are you going to fix these three commands?

Comment 6 Greg Bailey 2015-10-09 15:40:33 UTC
(In reply to Seiichi Ikarashi from comment #5)
> Greg, are you going to fix these three commands?

Yes, of course, I just returned from vacation and will address these either today or tomorrow.

Comment 7 Fedora Update System 2015-10-10 12:04:41 UTC
msr-tools-1.3-3.el7 has been submitted as an update to Fedora EPEL 7. https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2015-ace89f67cb

Comment 8 Fedora Update System 2015-10-12 18:50:11 UTC
msr-tools-1.3-3.el7 has been pushed to the Fedora EPEL 7 testing repository. If problems still persist, please make note of it in this bug report.
If you want to test the update, you can install it with
$ su -c 'yum --enablerepo=epel-testing update msr-tools'
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2015-ace89f67cb

Comment 9 Fedora Update System 2015-11-08 17:52:00 UTC
msr-tools-1.3-3.el7 has been pushed to the Fedora EPEL 7 stable repository. If problems still persist, please make note of it in this bug report.


Note You need to log in before you can comment on or make changes to this bug.