src: broadcast wall-velocity BC members in pre and post

bc_x/y/z%{vb1,vb2,vb3,ve1,ve2,ve3} are read from the namelist on rank 0
and consumed on all ranks by s_slip_wall/s_no_slip_wall in
src/common/m_boundary_common.fpp (~833-1155). These routines are compiled
into all three targets and are reached from pre_process (m_data_output,
m_perturbation) and post_process (m_start_up) code paths.

Without the broadcast, non-root ranks use uninitialised values for the
wall-velocity components, producing rank-dependent ghost cells for any
multi-rank wall-BC pre-process or post-process run. The sim residue has
always broadcast these; pre and post were missing them.

Add the matching broadcasts to both pre and post using the same nested
Fypp loop idiom as the simulation residue.

Author: sbryngelson

src/post_process/m_mpi_proxy.fpp

@@ -85,6 +85,14 @@ contains
             call MPI_BCAST(${VAR}$, 1, MPI_LOGICAL, 0, MPI_COMM_WORLD, ierr)
         #:endfor
 
+        ! wall-velocity members consumed by s_slip_wall/s_no_slip_wall on all ranks
+        #:for DIM in ['x', 'y', 'z']
+            #:for DIR in [1, 2, 3]
+                call MPI_BCAST(bc_${DIM}$%vb${DIR}$, 1, mpi_p, 0, MPI_COMM_WORLD, ierr)
+                call MPI_BCAST(bc_${DIM}$%ve${DIR}$, 1, mpi_p, 0, MPI_COMM_WORLD, ierr)
+            #:endfor
+        #:endfor
+
         ! manual: cfl_dt (runtime-computed logical), bc_io (BC-file existence)
         call MPI_BCAST(cfl_dt, 1, MPI_LOGICAL, 0, MPI_COMM_WORLD, ierr)
         call MPI_BCAST(bc_io, 1, MPI_LOGICAL, 0, MPI_COMM_WORLD, ierr)

src/pre_process/m_mpi_proxy.fpp

@@ -49,6 +49,14 @@ contains
             call MPI_BCAST(${VAR}$, 1, mpi_p, 0, MPI_COMM_WORLD, ierr)
         #:endfor
 
+        ! wall-velocity members consumed by s_slip_wall/s_no_slip_wall on all ranks
+        #:for DIM in ['x', 'y', 'z']
+            #:for DIR in [1, 2, 3]
+                call MPI_BCAST(bc_${DIM}$%vb${DIR}$, 1, mpi_p, 0, MPI_COMM_WORLD, ierr)
+                call MPI_BCAST(bc_${DIM}$%ve${DIR}$, 1, mpi_p, 0, MPI_COMM_WORLD, ierr)
+            #:endfor
+        #:endfor
+
         ! manual: cfl_dt (runtime-computed logical), bc_io (BC-file existence)
         call MPI_BCAST(cfl_dt, 1, MPI_LOGICAL, 0, MPI_COMM_WORLD, ierr)
         call MPI_BCAST(bc_io, 1, MPI_LOGICAL, 0, MPI_COMM_WORLD, ierr)