Memory scatter-gather mode is a complex mode used when data must be transferred to or from varied locations in memory instead of a set of contiguous locations in a memory buffer. For example, a gather μDMA operation could be used to selectively read the payload of several stored packets of a communication protocol, and store them together in sequence in a memory buffer.

In memory scatter-gather mode, the primary control structure is used to program the alternate control structure from a table in memory. The table is set up by the processor software and contains a list of control structures, each containing the source and destination end pointers, and the control word for a specific transfer. The mode of each control word must be set to memory scatter-gather mode. Each entry in the table is, in turn, copied to the alternate structure where it is then executed. The μDMA controller alternates between using the primary control structure to copy the next transfer instruction from the list, and then executing the new transfer instruction. The end of the list is marked by programming the control word for the last entry to use auto transfer mode. When the last transfer is performed using auto mode, the μDMA controller stops. A completion interrupt is generated only after the last transfer.

It is possible to loop the list by having the last entry copy the primary control structure to point back to the beginning of the list (or to a new list). It is also possible to trigger a set of other channels to perform a transfer, either directly, by programming a write to the software trigger for another channel, or indirectly, by causing a peripheral action that results in a μDMA request.

By programming the μDMA controller using this method, a set of arbitrary transfers can be performed based on a single μDMA request.

Figure 15-3 shows an example of operation in memory scatter-gather mode. This example shows a gather operation, where data in three separate buffers in memory is copied together into one buffer. Figure 15-3 shows how the application sets up a μDMA task list in memory, that is then used by the controller to perform three sets of copy operations from different locations in memory. The primary control structure for the channel used for the operation is configured to copy from the task list to the alternate control structure.

Figure 15-4 shows the sequence as the μDMA controller performs the three sets of copy operations. First, using the primary control structure, the μDMA controller loads the alternate control structure with Task A. The μDMA controller then performs the copy operation specified by Task A, copying the data from the source buffer A to the destination buffer. Next, the μDMA controller again uses the primary control structure to load Task B into the alternate control structure, and then performs the B operation with the alternate control structure. The process is repeated for Task C.

The application has a need to copy data items from three separate locations in memory into one combined buffer.

The application sets up µDMA "task list" in memory, which contains the pointers and control configuration for three µDMA copy "tasks."

The application sets up the channel primary control structure to copy each task configuration, one at a time, to the alternate control structure, where it is executed by the µDMA controller.

Figure 15-3 Memory Scatter-Gather, Setup, and Configuration

Figure 15-4 Memory Scatter-Gather, μDMA Copy Sequence