Hello everyone!

I've just sent v3 of part 1 and 2 of RFC series:
https://patchwork.kernel.org/patch/2836123/
https://patchwork.kernel.org/patch/2836124/


2013/7/17 Gerhard Sittig [off-list ref]:

OK, so the need for explicit start in software or external
request by the peripheral remains, but the condition for the
choice is different. It might become a flag attached to the DMA
channel, that gets setup in the prep routines (memory and slave
access each use their own prep calls) and gets evaluated in
execute.

I made a flag "will_access_peripheral" be set up in the device control callback.

non-MDDRC channels can get used for memory
transfers as well, which your description of the the former use
suggests, when slave support was absent.

I tested v3 on MPC5125 with SCLPC driver (transfers between dev and
mem work fine)
and dmatest module (all 64 DMA channels can perform mem-to-mem transfers).

quoted

quoted

+   mchan = dma_chan_to_mpc_dma_chan(chan);
+   switch (cmd) {
+   case DMA_TERMINATE_ALL:
+           /* disable channel requests */
+           mdma = dma_chan_to_mpc_dma(chan);
+           out_8(&mdma->regs->dmacerq, chan->chan_id);
+           list_splice_tail_init(&mchan->prepared, &mchan->free);
+           list_splice_tail_init(&mchan->queued, &mchan->free);
+           list_splice_tail_init(&mchan->active, &mchan->free);

This probably need locking.

Ah, yes.  It needs to grab the same lock as all the other list
manipulations in the driver's source.

Fixed that too.

Best regards,
Alexander.