Dave,
>
> > /* PaRAM slots are laid out like this */
> > struct edmacc_param {
> > unsigned int opt;
> > @@ -171,6 +176,10 @@ enum sync_dimension {
> > ABSYNC = 1
> > };
> >
> > +#define EDMA_CTLR_CHAN(ctlr, chan) (((ctlr) << 16) | (chan))
> > +#define EDMA_CTLR(i)
Dave,
[...]
>
> Plus, both your alloc_channel() and alloc_slot() are discarding the
> controller they may have been told to use ... alloc_slot() should
>
> if (slot >= 0) {
> ctlr = EDMA_CTLR(slot);
> slot = EDMA_CHAN(slot);
> }
I agree that I was discar
Dave,
I'll incorporate the comments and re-submit the patch.
Thanks, Sudhakar
>
> Quick reactions to this version ...
>
> Almost all of these changes flow directly from the interface
> changes which I've excerpted below, which will help make this
> response much shorter than the patch! That's
Quick reactions to this version ...
Almost all of these changes flow directly from the interface
changes which I've excerpted below, which will help make this
response much shorter than the patch! That's a nice feature
of a patch being reviewed.
Still ... 40 KBytes is big. This patch *might* sp
Existing EDMA driver does not take care of multiple EDMA channel
controller instances. An upcoming architecture from TI has 2 EDMA
channel controller instances. This patch adds support for multiple
EDMA CC instances.
With this patch, user needs to pass the EDMA channel number as
EDMA_CTLR_CHAN(ctl