This is the mail archive of the glibc-linux@ricardo.ecn.wfu.edu mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: sysconf(_SC_NPROCESSORS_CONF) core dump?


>>>>> Phil Estes writes:

 > Has anyone else encountered the subject mentioned problem under glibc 2.1.1
 > (RH 6.0)?  Under gdb the stack trace is quite confusing since sysconf calls
 > get_nprocs which then does something with the floppy drive (?!?) /dev/fd0,
 > which finally fails in a memcpy after some IO_ calls.  This also seems to
 > occur with _SC_NPROCESSORS_ONLN.  I was simply looking for a nice programmatic
This shouldn't happen - have a look at the code in
sysdeps/unix/sysv/linux/getsysstats.c.

 > way to determine number of processors and also whether it was uni- or multi-
 > processor mode.  I guess the other solution is to read /proc/cpuinfo--the
That's what glibc is doing.
 > reason I wasn't too excited about that was that I'm not sure if the format is
 > different on different kernel levels and whether someone will have a patch
 > which changes the format or uses a different process to show multiple cpus
 > (xosview recommends a patch that does something like this?).


 > Anyway, if I'm not using the glibc 2.1.1 sysconf() call correctly, can someone
 > give me a pointer.  And if it is a glibc bug, should I open a problem on this?
It's always good to give an example program  - otherwise I don't know
what you're doing and it save some time.

I've just created the following small program:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#include <unistd.h>
#include <stdio.h>

int
main (void)
{
  long r;
  
  r = sysconf (_SC_NPROCESSORS_CONF);
  printf ("_SC_NPROCESSORS_CONF is %d\n", r);
  r = sysconf (_SC_NPROCESSORS_ONLN);
  printf ("_SC_NPROCESSORS_ONLN is %d\n", r);

  return 0;
}
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Compiled it, and run it:
$ ./sysconf 
_SC_NPROCESSORS_CONF is 1
_SC_NPROCESSORS_ONLN is 1

No problems for me.  What are you doing differently?

If you feel this should be reported as a bug, please use the glibcbug
script and provide a way to reproduce it.

Andreas
-- 
 Andreas Jaeger   aj@arthur.rhein-neckar.de    jaeger@informatik.uni-kl.de
  for pgp-key finger ajaeger@aixd1.rhrk.uni-kl.de

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]