Re: [RFC PATCH 2/7] typec: tcpm: Add ADO header for Alert message handling

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Thu, Nov 02, 2017 at 05:34:22PM +0000, Adam Thomson wrote:
> On 02 November 2017 16:49, Greg Kroah-Hartman wrote:
> 
> > On Thu, Nov 02, 2017 at 11:40:12AM +0000, Adam Thomson wrote:
> > > On 01 November 2017 17:20, Greg Kroah-Hartman wrote:
> > >
> > > > On Wed, Nov 01, 2017 at 05:03:10PM +0000, Adam Thomson wrote:
> > > > > This commit adds a header providing definitions for handling Alert
> > > > > messages. Currently the header only focuses on handling incoming
> > > > > alerts.
> > > > >
> > > > > Signed-off-by: Adam Thomson <Adam.Thomson.Opensource@diasemi.com>
> > > > > ---
> > > > >  include/linux/usb/pd_ado.h | 49
> > > > ++++++++++++++++++++++++++++++++++++++++++++++
> > > > >  1 file changed, 49 insertions(+)
> > > > >  create mode 100644 include/linux/usb/pd_ado.h
> > > > >
> > > > > diff --git a/include/linux/usb/pd_ado.h b/include/linux/usb/pd_ado.h
> > > > > new file mode 100644
> > > > > index 0000000..edcbcfa
> > > > > --- /dev/null
> > > > > +++ b/include/linux/usb/pd_ado.h
> > > > > @@ -0,0 +1,49 @@
> > > > > +/*
> > > > > + * Copyright (c) 2017 Dialog Semiconductor
> > > > > + *
> > > > > + * Author: Adam Thomson <Adam.Thomson.Opensource@diasemi.com>
> > > > > + *
> > > > > + * This program is free software; you can redistribute it and/or modify
> > > > > + * it under the terms of the GNU General Public License as published by
> > > > > + * the Free Software Foundation; either version 2 of the License, or
> > > > > + * (at your option) any later version.
> > > >
> > > > I have to ask, do you _really_ mean "any later version"?
> > >
> > > Many thanks for the prompt comments.
> > >
> > > So this is the same copyright header applied for the other PD related code
> > > (bit of a copy and paste to follow existing tcpm practice), and have tended to
> > > use this header for other driver development as well. Within the context of the
> > > Linux kernel I believe this should be fine and will resolve to GPLv2, and should
> > > someone use the code elsewhere then they can choose a later GPL license, not
> > > that I suspect many would. Do you see any problems here as am happy to be
> > > corrected? :)
> >
> > Please consult your lawyers as to what license you should use for new
> > kernel code.  Never just blindly copy/paste without knowing what you are
> > supposed to be doing :)
> 
> Yes, our legal department are aware of the use of this header, from previous
> driver submissions, and have no problems. I certainly wouldn't have applied this
> otherwise, but completely understand why you raised this :)

Great, thanks for the info.

> > > > > + *
> > > > > + * This program is distributed in the hope that it will be useful,
> > > > > + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > > > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> > > > > + * GNU General Public License for more details.
> > > > > + */
> > > > > +
> > > > > +#ifndef __LINUX_USB_PD_ADO_H
> > > > > +#define __LINUX_USB_PD_ADO_H
> > > > > +
> > > > > +/* ADO : Alert Data Object */
> > > > > +#define ADO_TYPE_SHIFT			24
> > > > > +#define ADO_TYPE_MASK			0xff
> > > > > +#define ADO_FIXED_BATT_SHIFT		20
> > > > > +#define ADO_FIXED_BATT_MASK		0xf
> > > > > +#define ADO_HOT_SWAP_BATT_SHIFT		16
> > > > > +#define ADO_HOT_SWAP_BATT_MASK		0xf
> > > >
> > > > USB_ prefixes perhaps?
> > >
> > > Again here I was following the existing TCPM approach, but if it's a problem
> > > then I can update this.
> >
> > What does TCPM use here?  I missed those defines...
> 
> For example, include/linux/usb/pd_bdo.h uses defines such as:
> 
> #define BDO_MODE_TRANSMIT	(1 << 28)
> 
> and include/linux/usb/pd.h has defines such as:
> 
> #define PDO_FIXED_DUAL_ROLE	BIT(29)

Hm, ok, I guess that's ok for a define, but for a function:

> > > > > +#define ADO_TYPE_BATT_STATUS_CHANGE	BIT(1)
> > > > > +#define ADO_TYPE_OCP			BIT(2)
> > > > > +#define ADO_TYPE_OTP			BIT(3)
> > > > > +#define ADO_TYPE_OP_COND_CHANGE		BIT(4)
> > > > > +#define ADO_TYPE_SRC_INPUT_CHANGE	BIT(5)
> > > > > +#define ADO_TYPE_OVP			BIT(6)
> > > > > +
> > > > > +static inline unsigned int ado_type(u32 ado)
> > > > > +{
> > > > > +	return (ado >> ADO_TYPE_SHIFT) & ADO_TYPE_MASK;
> > > > > +}
> > > >
> > > > usb_ prefix for this and the other inline functions as well?
> > >
> > > Same comment here.
> >
> > But it would make more sense, right?
> 
> I'm happy with that. Actually I'd probably opt for 'usb_pd_' if I was going that
> route as it's PD specific information. Would that be reasonable?

Yes, that would be much better, thanks a lot.

greg k-h