Align dma_stream helper structure with mainline

[Je5s1e]

No description provided.

[hiroki-chen]

Please explain why is the snapshot for DmaStreamInner is justified:

• Why is better than wrapper method using rwarc locks?
• Who will provide the snapshot? Who tracks it? Who ensures that the snapshot is "the" current state of the dma stream?

To me this looks unsound.

[hiroki-chen]

Please explain why is the snapshot for DmaStreamInner is justified:

• Why is better than wrapper method using rwarc locks?
• Who will provide the snapshot? Who tracks it? Who ensures that the snapshot is "the" current state of the dma stream?

To me this looks unsound.

On a sidenote this design is intentional because tracking the runtime properties via RwArc is not well-supported. Maintaining the invariant without reasoning about how invariant mutates provides little verification meanings