[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDF #IKT-405664]: px_pgin, Write to unallocated



Hi Kevin,

> thanks for the suggestions. I've re-compiled my code and linked it
> against a slightly older version of netcdf libraries
> NETCDF-4.1.2-beta3-snapshot2011020522

Is there some reason you can't link against a more recent release,
such as 4.3.0, the current official release from May 2013, or
4.3.1-rc4, the current release candidate from Nov 2013?  If the 
bug is also in a current release, we can actually work on getting 
it fixed.  Also, to fix the bug, we need to be able to reproduce
the symptoms here, so we need a small demonstration program that
makes it less likely the bug is in user code, such as providing
an array argument not large enough for reading data into.

> this gave the following when running though dbx:
> 
> Write to unallocated (wua) on thread 18446744072618990672:
> Attempting to write 4 bytes through NULL pointer
> t@47591442046032 (l@27400) stopped in NC_initialize at line 36 in file
> "stub.c"
> 36       if((stat = NC4_initialize())) return stat;

Sorry, but that's not enough information to determine the cause of the 
problem, so there's not much we can do with it, unless we can reproduce 
it.  It would be helpful if you observed such a bug running any of the
tests run by "make check", in which case we would have a chance of
isolating the problem to the test program or the library, and could
diagnose and fix the cause ...

> I continued and got:
> 
> Write to unallocated (wua) on thread 18446744072618990672:
> Attempting to write 4 bytes at address 0x569178c4d280
> t@47591442046032 (l@27400) stopped in NC_dispatch_overlay at line 157 in
> file "dispatch.c"
> 157       vmerge = (void**)merge;
> 
> interestingly if I try and print these variables I get:
> 
> print vmerge
> dbx: warning: variable `vmerge' has no address - unused or optimized out
> (dbx) print merge
> merge = 0x2b48bc626940
> 
> so is the problem in compiling the netcdf with optimization on. I
> compiled without specifically setting optimization so don't know if it
> was chosen by default. Also the netcdf libs are compiled with gcc whilst
> my code has been compiled with Sun studio cc and f95.
> 
> I can try and re-compile the hdf5 and netcdf-4 with Sun studio
> compilers.

If it's convenient, could you also try using the current releases or 
release candidates for HDF5 and netCDF-4?

For netCDF-4, these are:
   ftp://ftp.unidata.ucar.edu/pub/netcdf/netcdf-4.3.0.tar.gz
   ftp://ftp.unidata.ucar.edu/pub/netcdf/netcdf-4.3.1-rc4.tar.gz
with build instructions here:
   http://www.unidata.ucar.edu/software/netcdf/docs/building.html

--Russ

> Regards
> 
> Kevin
> --
> ---------------------------------------------------------
> |                                                        |
> | Kevin I.Hodges                                         |
> | ESSC (Enviromental Systems Science Center)             |
> | University of Reading,                                 |
> | Harry Pitt Building,                                   |
> | 3 EarleyGate,                                          |
> | Whiteknights, PO Box 238,                              |
> | Reading RG6 6AL, U.K.                                  |
> |                                                        |
> | Tel:  0118 3787761                                     |
> | Fax:  0118 3786413                                     |
> |                                                        |
> | email: address@hidden                        |
> | WWW: http://www.nerc-essc.ac.uk/~kih/Home.html         |
> |                                                        |
> ---------------------------------------------------------
> 
> 
Russ Rew                                         UCAR Unidata Program
address@hidden                      http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: IKT-405664
Department: Support netCDF
Priority: Normal
Status: Closed