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

From: Greg KH
Date: Thu Dec 12 2019 - 12:32:56 EST


On Thu, Dec 12, 2019 at 05:25:28PM +0000, Vitor Soares wrote:
> From: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx>
> Date: Thu, Dec 12, 2019 at 16:00:45
>
> > On Thu, Dec 12, 2019 at 02:56:56PM +0000, Vitor Soares wrote:
> > > Hi Greg,
> > >
> > > From: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx>
> > > Date: Thu, Dec 12, 2019 at 14:44:59
> > >
> > > > On Tue, Dec 10, 2019 at 04:37:33PM +0100, Vitor Soares wrote:
> > > > > +static int __init i3cdev_init(void)
> > > > > +{
> > > > > + int res;
> > > > > +
> > > > > + pr_info("i3c /dev entries driver\n");
> > > >
> > > > Please remove debugging information, kernel code should be quiet unless
> > > > something goes wrong.
> > >
> > > I will remove it.
> > >
> > > >
> > > > > + /* Dynamically request unused major number */
> > > > > + res = alloc_chrdev_region(&i3cdev_number, 0, N_I3C_MINORS, "i3c");
> > > >
> > > > Do you really need a whole major, or will a few minors work?
> > > >
> > > > thanks,
> > > >
> > > > greg k-h
> > >
> > > I'm reserving one per device. What do you suggest?
> >
> > How many devices do you have in a system?
>
> According to MIPI I3C spec, the maximum number of devices on a bus will
> depend on trace length, bus load, and clock requirements.
>
> Frankly, I don't know what is the best compromise because it depends from
> system to system and the end-use of it. In my case, I started developing
> this to help me during the HC controller driver development.
> Even If I choose a fixed major, I wouldn't know in which one i3c fits.

Ok, a full major will be fine, I was worried you only wanted 1 or 2
minor numbers, which would mean you could have used a misc device
instead.

This is fine.

thanks,

greg k-h