[petsc-dev] Segfault in DMPlexVecGetClosure

[Lawrence Mitchell]

> On 31 Oct 2017, at 15:36, Blaise A Bourdin <bourdin at lsu.edu> wrote:
> 
> Hi,
> 
> I am losing my mind over segfaults in DMPlexVecGetClosure.
> 
> In the example below, I get a segfault whenever cval is not set to NULL before calling DMPlexVecGetClosure. Is this expected? The documentations says nothing about this
> Also, after calling DMPlexVecRestoreClosure, the values of cval are discarded and neither global nor coord are changed.
> 
> I am obviously doing something wrong but cannot figure it out...

The interface is similar to DMPlexGetClosure.

DMPlexVecGetClosure fills the array you provide it with values.  It can allocate a work array for you if you don't provide one, but you must indicate so explicitly by setting it to NULL:

so

PetscScalar *values = NULL;

DMPlexVecGetClosure(...., &values);

Do things with values

Now you need to release the work array, you do that with the matching restore closure call.

DMPlexVecRestoreClosure(...., &values);


Now, these arrays you get are *not* pointers into the global Vec array, instead they are local buffers.  If you want to update the Vec with the values you've put in values, you need to call DMPlexVecSetClosure:

e.g.

DMPlexVecSetClosure(..., values, INSERT_VALUES);


So the calling sequence should be something like:

PetscScalar *values = NULL;

for ( int p = 0; p < ...; p++ ) {
    DMPlexVecGetClosure(..., p, ..., &values);
    // do stuff with values.
    DMPlexVecSetClosure(..., p, ..., values, INSERT_VALUES);
}
// Finally, restore work array
DMPlexVecRestoreClosure(..., 0, ..., &values);

Cheers,

Lawrence