Bug 952422 – sysconf(_SC_NPROCESSORS_ONLN) performance problem

Bugzilla will be upgraded to version 5.0. The upgrade date is tentatively scheduled for 2 December 2018, pending final testing and feedback.

Bug 952422 - sysconf(_SC_NPROCESSORS_ONLN) performance problem

Summary:	sysconf(_SC_NPROCESSORS_ONLN) performance problem

Status:	CLOSED ERRATA

Aliases:	None

Product:	Red Hat Enterprise Linux 6
Classification:	Red Hat
Component:	glibc (Show other bugs)
Sub Component:
Version:	6.5
Hardware:	All Linux

Priority	high Severity high
Target Milestone:	rc
Target Release:	---
Assigned To:	Siddhesh Poyarekar
QA Contact:	Arjun Shankar
Docs Contact:

URL:
Whiteboard:
Keywords:	Patch

Depends On:
Blocks:
	Show dependency tree / graph

Reported:	2013-04-15 19:19 EDT by Jacob Hunt
Modified:	2016-11-24 07:14 EST (History)
CC List:	7 users (show)

See Also:
Fixed In Version:
Doc Type:	Bug Fix
Doc Text:	This update improves performance of repeated calls of the sysconf function for getting values of _SC_NPROCESSORS_ONLN by caching the value for approximately a second. This could improve performance of some applications that use QReadWriteLocks in the Qt library or any other functionality that involves repeatedly getting the value for _SC_NPROCESSORS_ONLN.
Story Points:	---
Clone Of:
Environment:
Last Closed:	2013-11-21 05:45:08 EST
Type:	Bug
Regression:	---
Mount Type:	---
Documentation:	---
CRM:
Verified Versions:
Category:	---
oVirt Team:	---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---

Attachments	(Terms of Use)
Candidate fix (3.04 KB, patch) 2013-05-29 02:01 EDT, Siddhesh Poyarekar	no flags	Details \| Diff
Add an attachment (proposed patch, testcase, etc.)

External Trackers
Tracker	ID	Priority	Status	Summary	Last Updated
Red Hat Product Errata	RHSA-2013:1605	normal	SHIPPED_LIVE	Moderate: glibc security, bug fix, and enhancement update	2013-11-20 16:54:09 EST

Groups: None (edit)

Description Jacob Hunt 2013-04-15 19:19:45 EDT

Description of problem:

We have an application that has a lot of threads and uses QReadWriteLocks to protect data that the threads are accessing. When our server is heavily loaded, the application spends about 60% of it's time in system time as opposed to user time. Further investigation showed that the QMutex class is calling sysconf(_SC_NPROCESSORS_ONLN) for each thread that tries to get a lock where that lock is already locked. Thus most of that system time is spent in this system call.

Version-Release number of selected component (if applicable):
glibc-2.12-1.107.el6.x86_64


Additional info:
This has been discussed upstream: 
http://sourceware.org/ml/libc-alpha/2011-06/msg00064.html

Comment 11 Siddhesh Poyarekar 2013-05-29 02:01:42 EDT

Created attachment 754181 [details]
Candidate fix

Comment 14 errata-xmlrpc 2013-11-21 05:45:08 EST

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHSA-2013-1605.html

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