Re: [RFC 5/5] i3c: add i3cdev module to expose i3c dev in /dev

[Arnd Bergmann <arnd@arndb.de>]

On Wed, Dec 11, 2019 at 4:07 PM Vitor Soares <Vitor.Soares@synopsys.com> wrote:
> From: Arnd Bergmann <arnd@arndb.de>
> Date: Tue, Dec 10, 2019 at 19:37:14
> > On Tue, Dec 10, 2019 at 8:15 PM Vitor Soares <Vitor.Soares@synopsys.com> wrote:
> > > From: Arnd Bergmann <arnd@arndb.de>
> > > Date: Tue, Dec 10, 2019 at 17:51:14
> > > > On Tue, Dec 10, 2019 at 4:37 PM Vitor Soares <Vitor.Soares@synopsys.com> wrote:
> > As your interface is basically just read/write based, I wonder
> > if there is a way to completely avoid the ioctl and instead
> > use io_submit() as the primary interface.
>
> I confess that I wasn't familiar with io_submit() until now and went
> straightway for the ioctl() approach.
> So far, my understanding is that io_submit() will call .write or .read of
> i3cdev module depending on the iocb command. if so, we won't be able to
> do a repeated start between a multiple iocb in the same list, right?

I'm not sure what you mean with "repeated start", but it's definitely
possible that io_submit() is not a useful interface for i3c. The main
advantage would be that it avoids creating a complex ioctl command.

> Apart from this private read/write need, another requirement that leads
> me to use the ioctl() was:
> - When we support HDR command in i3c subsystem we can expand the ioctl()
> command to support it.
> - For now, device API doesn't expose CCC commands but some of them are
> used for a private contract between master and device and we likely need
> that support in the future as well.

I think you could still have both the io_submit() interface for basic I/O
(if you can get it to do what you want), plus an ioctl interface for more
complex interactions.

     Arnd