On Tue, Jul 12, 2011 at 6:17 AM, Jassi Brar [off-list ref] wrote:

1) Striding, in one form or other, is supported by other DMACs as well.
? The number will only increase in future.
? Are we to add ?<VENDOR>_DMA_STRIDE_CONFIG for each case ?

If we are sure about this and striding will work in a similar way on all
then let's have the enum named DMA_STRIDE_CONFIG and move the
passed-in struct to <linux/dmaengine.h) then?

Would that be:

struct dma_stride_config {
    u32 read_bytes;
    u32 skip_bytes;
};

Or something more complex?

2) As Dan noted, client drivers are going to have ifdef hackery in
order to be common
?to other SoCs.

Don't think so, why? This is a runtime config entirely, and I just illustrated
in mail to Dan how that can be handled by falling back to a sglist I believe?

We can *maybe* even put the fallback code into dmaengine, so that an
emulated sglist in place for the DMAengine is done automatically of the
DMA controller does not support striding.

3) TI may not have just one DMAC IP used in all the SoCs. So if you want
?vendor specific defines anyway, please atleast also add DMAC version to it.
?Something like

quoted

? ? ? ?DMA_SLAVE_CONFIG,
? ? ? ?FSLDMA_EXTERNAL_START,
+ ? ? ? TI_DMA_v1_STRIDE_CONFIG,

Yep unless we make it generic DMA_STRIDE_CONFIG simply, this makes
a lot of sense.

Linus Walleij