[Mauiusers] Bug in MRes.c in maui 3.2.6p14

Scott Jackson scottmo at clusterresources.com
Thu Sep 1 10:09:19 MDT 2005


Thomas,

This issue was cosmetic in nature and has been corrected in the latest
snapshot release of Maui. Thank you for your submission.

Scott Jackson
Cluster Resources, Inc.

 
On Thu, 2005-09-01 at 07:26, David B Jackson wrote:
> Thomas,
> 
>   Thanks!  We will get out a patched snapshot this morning.
> 
> Dave
> 
> > I investigated this file after trying to find the source of a maui
> > crash with valgrind. Valgrind detected some invalid read of size 1 on
> > the following piece of code, starting at line 4303 :
> >
> >     for (aindex = 0;aindex < MAX_MACL;aindex++) {
> >       if (J->Cred.CL[aindex].Type == maDuration) {
> >           OTime = J->Cred.CL[aindex].Value;
> >           J->Cred.CL[aindex].Value = Overlap;
> >           break;
> >       }
> >     }
> >
> >     if (J->Cred.CL[aindex].Type == maNONE) {   <---- valgrind error here
> >       ....
> >
> > The problem shown by valgrind is simple : in some cases the for loop
> > goes to aindex== MAX_MACL. After that all the code is broken because
> > we are working on memory that is not even part of the array (confirmed
> > by valgrind). I'm submitting this to you because for sure this piece
> > of code is buggy, but I have absolutly no idea of how to fix it.
> >
> > I am still analysing other valgrind traces to find the double free bug
> > that is killing our scheduler. I will probably submit more problems
> > like this one in the near future.
> >
> > Regards,
> > Thomas.
> > _______________________________________________
> > mauiusers mailing list
> > mauiusers at supercluster.org
> > http://www.supercluster.org/mailman/listinfo/mauiusers
> >
> 
> _______________________________________________
> mauiusers mailing list
> mauiusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/mauiusers



More information about the mauiusers mailing list