Re: [PATCH] spi: spi-mem: add statistics support to ->exec_op() calls

[Mark Brown]

On Mon, Feb 12, 2024 at 04:43:55PM +0530, Dhruva Gole wrote:
> On Feb 09, 2024 at 14:51:23 +0100, Théo Lebrun wrote:

> > +	/*
> > +	 * We do not have the concept of messages or transfers. Let's consider
> > +	 * that one operation is equivalent to one message and one transfer.

> Why 1 message _and_ 1 xfer and not simply 1 xfer?
> Even in the example of testing that you showed above the values for
> message and xfer are anyway going to be same, then why have these 2
> members in the first place? Can we not do away with one of these?

If the device supports regular SPI operations as well as spi-mem
operations then this will ensure that the spi-mem stats fit in with the
other operations.  If it only supports spi-mem operations it would not
matter so much.

Attachment: signature.asc
Description: PGP signature