Re: [PATCH v3] spi: core: Validate length of the transfers in message

[Atsushi Nemoto <anemo@mba.ocn.ne.jp>]

On Wed, 26 Feb 2014 09:14:07 +0900, Mark Brown <broonie@kernel.org> wrote:
>>       + optionally defining short delays after transfers ... using
>>         the spi_transfer.delay_usecs setting (this delay can be the
>>         only protocol effect, if the buffer length is zero);
> 
> *sigh*  I guess it is valid, though frankly I'm concerned that this
> isn't a good idea - it's certainly not going to work reliably given the
> need for every driver to open code this, most of them get the delay
> stuff wrong.

I don't object to the whole patch.  Validating in spi core is good of
course, and "xfer->len % w_size" part looks no problem.

I just want to keep ways to handle an odd device, for example, which
requires long delay between chipselect and the first transfer, etc.

---
Atsushi Nemoto