[LowerToAIE] Add support for DMA chains

[jtuyls]

Adds support for lowering amdaie.npu.circular_dma_cpy_nd to a chain of aie.dma_bd operations. This is needed becaused low-level DMA BD configurations currently don't support a zero stride.

For example:

amdaie.npu.circular_dma_cpy_nd %some_connection([] [] [], [0, 0, 0] [2, 32, 32] [0, 64, 1])

contains an outer dimension with zero stride, which describes a repetition, and if this is not the common repetition count of all connections operating on a logical objectFifo, this is lowered to a chain of dma_bd operations to implement the zero stride:

^bb3:  // 2 preds: ^bb2, ^bb4
  aie.use_lock(%lock_0_1_0, AcquireGreaterEqual, 1)
  aie.dma_bd(%buffer_0_1 : memref<4096xi32, 1 : i32>) {dimensions = #aie<bd_dim_layout_array[<size = 32, stride = 64>, <size = 32, stride = 1>]>, len = 1024 : i32}
  aie.next_bd ^bb4
^bb4:  // pred: ^bb3
  aie.dma_bd(%buffer_0_1 : memref<4096xi32, 1 : i32>) {dimensions = #aie<bd_dim_layout_array[<size = 32, stride = 64>, <size = 32, stride = 1>]>, len = 1024 : i32}
  aie.use_lock(%lock_0_1, Release, 1)
  aie.next_bd ^bb3

Note how the first block contains a lock acquire operation, but no lock release operation as it needs to hand off to the second block before releasing a lock to create the DMA chain.

[Yu-Zhewen]

LGTM, just one nit