Re: [PATCH v4 0/6] spi: spi-fsl-dspi: Target mode improvements

[Mark Brown]

On Tue, Jul 01, 2025 at 06:24:33PM +0300, Vladimir Oltean wrote:
> On Tue, Jul 01, 2025 at 04:16:50PM +0100, Mark Brown wrote:

> > Switching between modes is incredibly common, usually between PIO (for
> > very short transfers) and DMA, that's no problem.  Factoring in
> > timestamping seems like a reasonable signal I guess, might trip someone
> > who was trying to benchmark things up but probably not normal users.

> Ah, ok, I vaguely remember something being discussed about can_dma()
> on previous iterations of this patch, but in a different context.
> Then that's an avenue to explore, I guess. Looking at that method's
> prototype, I suppose dspi could simply return can_dma = false "if (xfer->ptp_sts)"
> (timestamp requested), i.e. no core involvement in the decision process at all?

Yes, exactly.  It can base the decision on whatever amuses it about the
transfer.

Attachment: signature.asc
Description: PGP signature