On 12/15/2014 12:40 PM, Eric Chamberland wrote:

Hi,

I just reported this to OpenMPI (http://www.open-mpi.org/community/lists/users/2014/12/25997.php), but when trying with MPICH 3.1.3, valgrind reports problems too.

Someone kindly proposed a patch, but since the code seems to be shared by both projects, I thought I should report it here too (or should I not?)

I check in with the open-mpi list from time to time, but I pay more attention to mpich-discuss. No harm in letting both lists know about problems you find.

With valgrind, we detect:

1- two "Invalid read of size 1" 2- Warning: invalid file descriptor -1 in syscall read() 3- Syscall param writev(vector[...]) points to uninitialised byte(s)

Setup to make this happens:

thanks for the test case. Should be an easy fix, but I cannot get to it right this moment: https://trac.mpich.org/projects/mpich/ticket/2212 ==rob

export too_long=./this/is/a_very/long/path/that/contains/a/not/so/long/filename/but/trying/to/collectively/mpi_file_open/it/you/will/have/a/memory/corruption/resulting/of/invalide/writing/or/reading/past/the/end/of/one/or/some/hidden/strings/in/mpio/Simple/user/would/like/to/have/the/parameter/checked/and/an/error/returned/or/this/limit/removed

mkdir -p $too_long echo "hello world." > $too_long/toto.txt

mpicc -o bug_MPI_File_open_path_too_long bug_MPI_File_open_path_too_long.c

mpirun -np 2 valgrind ./bug_MPI_File_open_path_too_long $too_long/toto.txt

This particular simple case doesn't segfault without valgrind, but in as reported in my real case, it does!

Thanks!

Eric

_______________________________________________ discuss mailing list [email protected] To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

-- Rob Latham Mathematics and Computer Science Division Argonne National Lab, IL USA _______________________________________________ discuss mailing list [email protected] To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss