Manta Interactive Ray Tracer Development Mailing List

Text archives Help


Re: [MANTA] bad exit from bin/manta on Altix


Chronological Thread 
  • From: James Bigler <bigler@cs.utah.edu>
  • Cc: manta@sci.utah.edu
  • Subject: Re: [MANTA] bad exit from bin/manta on Altix
  • Date: Wed, 25 May 2005 10:54:45 -0600
linuxthreads

[bigler@muzzle ~/manta/linuxopt]% setenv LD_ASSUME_KERNEL 2.4.19
[bigler@muzzle ~/manta/linuxopt]% ldd bin/manta | grep pthread
        libpthread.so.0 => /lib/i686/libpthread.so.0 (0x40137000)
[bigler@muzzle ~/manta/linuxopt]% ldd bin/manta | grep libc
        libc.so.6 => /lib/i686/libc.so.6 (0x4047f000)
[bigler@muzzle ~/manta/linuxopt]% getconf GNU_LIBPTHREAD_VERSION
linuxthreads-0.10

NTPL

[bigler@muzzle ~/manta/linuxopt]% unsetenv LD_ASSUME_KERNEL 2.4.19
[bigler@muzzle ~/manta/linuxopt]% ldd bin/manta | grep pthread
        libpthread.so.0 => /lib/tls/libpthread.so.0 (0x40137000)
[bigler@muzzle ~/manta/linuxopt]% ldd bin/manta | grep libc
        libc.so.6 => /lib/tls/libc.so.6 (0x4043e000)

---------------------------------------------------------------
Inspecting the libc libraries like you mentioned results in this:

[bigler@muzzle ~/manta/linuxopt]% /lib/i686/libc.so.6
GNU C Library stable release version 2.3.3, by Roland McGrath et al.
Copyright (C) 2003 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 3.3.2 (Mandrake Linux 10.0 3.3.2-4mdk).
Compiled on a Linux 2.6.0 system on 2004-02-16.
Available extensions:
        GNU libio by Per Bothner
        crypt add-on version 2.1 by Michael Glad and others
        linuxthreads-0.10 by Xavier Leroy
        BIND-8.2.3-T5B
        libthread_db work sponsored by Alpha Processor Inc
        NIS(YP)/NIS+ NSS modules 0.19 by Thorsten Kukuk
Thread-local storage support included.
Report bugs using the `glibcbug' script to <bugs@gnu.org>.


[bigler@muzzle ~/manta/linuxopt]% /lib/tls/libc.so.6
GNU C Library stable release version 2.3.3, by Roland McGrath et al.
Copyright (C) 2003 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 3.3.2 (Mandrake Linux 10.0 3.3.2-4mdk).
Compiled on a Linux 2.6.0 system on 2004-02-16.
Available extensions:
        GNU libio by Per Bothner
        crypt add-on version 2.1 by Michael Glad and others
        NPTL 0.60 by Ulrich Drepper
        BIND-8.2.3-T5B
        NIS(YP)/NIS+ NSS modules 0.19 by Thorsten Kukuk
Thread-local storage support included.
Report bugs using the `glibcbug' script to <bugs@gnu.org>.


---------------------------------------------------------------
So it would appear that I'm using NTPL.

James

Thiago Ize wrote:
What do you get when you type:

/lib/libc.so.*


If it says "linuxthreads-0.9 by Xavier Leroy" (or something similar), then you aren't using the NPTL which is "Native POSIX Threads Library by Ulrich Drepper et al."

Here's where I got my information: http://gentoo-wiki.com/NPTL

Also, fisher and prism, which are our altix machines, run 2.4 kernels, so we don't get that problem (we also don't get all the 2.6 kernel benifits...). Supposedly NPTL is faster than linuxthreads, so getting manta to work correctly with it is a good thing.

Thiago

James Bigler wrote:

So this is the kernel I'm using (stock from Mandrake 10.0), and I don't see the errors you report. I'm using a fresh update as well.

Linux muzzle.sci.utah.edu 2.6.3-7mdk-p3-smp-64GB #1 SMP Wed Mar 17 15:34:39 CET 2004 i686 unknown unknown GNU/Linux

Perhaps my kernel isn't new enough?

James

Rocky Rhodes wrote:

It will show up in other linux distributions.  Google'ing "nptl
linuxthreads" gave me a bunch of hits including:

http://linuxdevices.com/articles/AT6753699732.html

I had tried older versions of the main trunk and didn't have any luck
isolating a change. I tried versions back to 308 that all failed similarly,
and version 300 doesn't build for me. It has probably always been this way
and I just had this environment variable set when I tried it before.

    Rocky


-----Original Message-----
From: owner-manta@sci.utah.edu [mailto:owner-manta@sci.utah.edu] On Behalf
Of James Bigler
Sent: Tuesday, May 24, 2005 7:59 PM
Cc: manta@sci.utah.edu
Subject: Re: [MANTA] bad exit from bin/manta on Altix

You could try checking out an older version and see if it happens.

The Thread_pthread.cc file is pretty hairy.  The altix is the only
machine we've had complaints about, though.  Should this be showing up
in other modern distributions?  What do the SGI docs say about that?

James

Rocky Rhodes wrote:

If I run "bin/manta -bench 10 10 -imagedisplay null -np 2" on an Altix,
the program exits in a bad way, complaining of "ERROR: Incorrect Phase"
and then telling me that "Thread 'idle or main'" got a SIGSEGV.  If I
run this again with the LD_ASSUME_KERNEL environment variable set to
"2.4.19" it exits cleanly.



SGI's documentation says that this behavior is indicative of an
application "which depends on behaviors in which the LinuxThreads
implementation deviates from the POSIX standard". The LD_ASSUME_KERNEL
environment variable forces the application to use the old LinuxThreads
implementation rather than NPTL (Native POSIX Thread Library). I think
the new thread package was included with ProPack 3.0 on the Altix. You
might not see this problem on your Altix if it is running an earlier
version of the system software.



I thought I had tried this earlier and didn't have this problem, but as
it is just an environment variable, now I'm wondering if this has always
been broken this way.  Is anyone aware of any changes in the pthread
code made over the last week or so that may have changed this behavior?
Does anyone feel more qualified than I do about mucking around in this
code and trying to understand what goes on when the program exits?



           Rocky








Archive powered by MHonArc 2.6.16.

Top of page