Bug 121

Summary: get_nprocs and get_nprocs_conf do not seem to be complete...
Product: uClibc Reporter: Ryan Hope <rmh3093>
Component: Standard ComplianceAssignee: unassigned
Status: RESOLVED INVALID    
Severity: enhancement CC: rep.dot.nop, uclibc-cvs, vda.linux
Priority: P5    
Version: 0.9.30   
Target Milestone: ---   
Hardware: PC   
OS: Linux   
Host: Target:
Build:

Description Ryan Hope 2009-02-19 16:12:44 UTC
get_nprocs and get_nprocs_conf do not seem to be complete...
in /usr/include/sys/sysinfo.h there are void stubs which throw a warning/error

maybe sysconf(_SC_NPROCESSORS_CONF) and sysconf(_SC_NPROCESSORS_ONLN) could be used for get_nprocs_conf and get_nprocs respectively
Comment 1 Ryan Hope 2009-02-19 17:00:10 UTC
ehh i think i was wrong in my last post, num_online_cpus() and int num_present_cpus() in include/linux/cpumask.h is what needs to be used probably
Comment 2 Denys Vlasenko 2009-03-02 21:06:04 UTC
I do not understand this bug report.

What are you trying to do? Do you have a code example which works with e.g. glibc but fails with uclibc?
Comment 3 Ryan Hope 2009-03-02 22:40:35 UTC
(In reply to comment #2)
> I do not understand this bug report.
> 
> What are you trying to do? Do you have a code example which works with e.g.
> glibc but fails with uclibc?
> 

I think it was boost that I was trying to compile when I got this error...

./libs/thread/src/pthread/thread.cpp
Comment 4 Bernhard 2009-03-02 22:44:11 UTC
I didn't find the time yet but i fear that this hunk is not adequate, is it?
Thanks in advance for checking and reporting back!
Comment 6 Mike Frysinger 2009-03-16 05:35:23 UTC
this is a bug in boost.  it uses glibc-specific stuff instead of the POSIX compliant method.  punt this bug as invalid.

https://svn.boost.org/trac/boost/ticket/2053
Comment 7 Mike Frysinger 2009-03-16 05:39:48 UTC
which is not to say that the patch referenced by Bernhand shouldnt be added (it should be).  just that adding the symbols get_nprocs/get_nprocs_conf is incorrect.
Comment 8 Bernhard Reutner-Fischer 2010-06-22 20:20:36 UTC
CFT patch: http://lists.uclibc.org/pipermail/uclibc/2010-April/043814.html