Allow user to specify CPU VA for device allocation via dxgkio_lock2()

Nobody has claimed this yet.

**Is your feature request related to a problem? Please describe.**
NO

**Describe the solution you'd like**
The motivation is that we want to use unified VA for GPU and CPU for a device allocation.

Use case:
1, Create a device allocation with dxgkio_create_allocation()
2, Map the allocation and get a GPU VA with dxgkio_map_gpu_va()
3, Get a CPU VA of the allocation with dxgkio_lock2()

The problem is GPU VA != CPU VA with dxgkio_lock2() implementation.

struct d3dkmt_lock2 {
struct d3dkmthandle device;
struct d3dkmthandle allocation;
struct d3dddicb_lock2flags flags;
__u32 reserved;
#ifdef __KERNEL__
void *data;
#else
__u64 data;
#endif
};

The solution is allowing user to specify VA through data filed.
It would be better to specify mapping flags(read/write/execute) through flags filed.

**Describe alternatives you've considered**

**Additional context**
https://docs.nvidia.com/cuda/cuda-driver-api/group__CUDA__UNIFIED.html

Start by tracing the dxgkio_lock2() path and comparing it with dxgkio_create_allocation() and dxgkio_map_gpu_va(). Review the d3dkmt_lock2 structure and the linked CUDA unified-addressing documentation. Done means the allocation can use a user-specified CPU VA matching the GPU VA, with the requested mapping permissions represented by the flags.

Written by the indexing model from the issue text.