http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51073
Bug #: 51073
Summary: _gfortran_caf_register incorrectly assumes malloc(0)
returns non-NULL
Classification: Unclassified
Product: gcc
Version: 4.7.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
AssignedTo: [email protected]
ReportedBy: [email protected]
Revision: Wed Nov 9 13:24:05 UTC 2011 (revision 181203)
Target: mips-rtems4.11 but impacts all where malloc(0) returns NULL
I am debugging why some of the gfortran tests are failing. I have tracked NN
failures down to this code in caf/mpi.c around line 155.
/* Token contains only a list of pointers. */
local = malloc (size);
token = malloc (sizeof (void*) * caf_num_images);
if (unlikely (local == NULL || token == NULL))
goto error;
>From http://pubs.opengroup.org/onlinepubs/009695399/functions/malloc.html
Upon successful completion with size not equal to 0, malloc() shall return a
pointer to the allocated space. If size is 0, either a null pointer or a unique
pointer that can be successfully passed to free() shall be returned. Otherwise,
it shall return a null pointer [CX] and set errno to indicate the error.
The code uses the memory pointed to by local to pass to MPI_Allgather() and
then returns it to the caller. I suspect that the line "local = malloc (size)"
should be something like:
local = malloc ((size == 0)? sizeof(XXX) : size);
Where XXX needs to be the type of one entity of sufficient size to pass into
MPI_Allgather.
