Trying to force Eigen matrices to hardcoded addresses - eigen

Question

I've got a large Eigen-based library that runs great on Linux, but now needs to be ported to an embedded environment without a heap.
In theory this seems doable using Map, but I'd appreciate if someone could confirm my understanding of the specifics.
All the Eigen matrices are currently stored in a struct, e.g.,
struct {
Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> array1;
Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> array2;
} my_struct;
my_struct.array1 = Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>::Zero(rows, cols);
my_struct.array2 = Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>::Zero(rows, cols);
I tried using Map in the struct but that gives me assorted compile errors and I think I'm probably looking at the problem the wrong way.
So now I'm thinking I should do something like
struct {
double *array1_addr;
double *array2_addr;
} my_struct;
my_struct.array1_addr = (double *)0x100000; // address of first array
my_struct.array2_addr = (double *)0x200000; // address of second array
Eigen::Map< Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> > map1(my_struct.array1_addr, rows, cols);
map1.setZero();
Eigen::Map< Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic> > map2(my_struct.array2_addr, rows, cols);
map2.setZero();
That is, I take the arrays all out of the struct and simply use Map to respecify the address of the array wherever I use it.
Is there a better approach?