Re: [PATCH 1/6] HID: intel-ish-hid: add support for MODULE_DEVICE_TABLE()

[Thomas Weißschuh]

On 2021-11-01 10:58+0100, Hans de Goede wrote:
> On 10/29/21 17:28, Thomas Weißschuh wrote:
> > This allows to selectively autoload drivers for ISH devices.
> > Currently all ISH drivers are loaded for all systems having any ISH
> > device.
> > 
> > Signed-off-by: Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
> > 
> > ---
> > 
> > Cc: linux-kbuild@xxxxxxxxxxxxxxx
> > Cc: linux-input@xxxxxxxxxxxxxxx
> > Cc: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > Cc: Jiri Kosina <jkosina@xxxxxxx>
> > Cc: Benjamin Tissoires <benjamin.tissoires@xxxxxxxxxx>
> > Cc: Hans de Goede <hdegoede@xxxxxxxxxx>
> > Cc: Masahiro Yamada <masahiroy@xxxxxxxxxx>
> > Cc: Michal Marek <michal.lkml@xxxxxxxxxxx>
> > Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
> > ---
> >  include/linux/mod_devicetable.h   | 13 +++++++++++++
> >  scripts/mod/devicetable-offsets.c |  3 +++
> >  scripts/mod/file2alias.c          | 24 ++++++++++++++++++++++++
> >  3 files changed, 40 insertions(+)
> > 
> > diff --git a/include/linux/mod_devicetable.h b/include/linux/mod_devicetable.h
> > index ae2e75d15b21..befbf53c4b7c 100644
> > --- a/include/linux/mod_devicetable.h
> > +++ b/include/linux/mod_devicetable.h
> > @@ -895,4 +895,17 @@ struct dfl_device_id {
> >  	kernel_ulong_t driver_data;
> >  };
> >  
> > +/* ISHTP (Integrated Sensor Hub Transport Protocol) */
> > +
> > +#define ISHTP_MODULE_PREFIX	"ishtp:"
> > +
> > +/**
> > + * struct ishtp_device_id - ISHTP device identifier
> > + * @guid_string: 36 char string of the form fa50ff2b-f2e8-45de-83fa-65417f2f49ba
> > + * @context: pointer to driver specific data
> > + */
> > +struct ishtp_device_id {
> > +	guid_t guid;
> 
> The kdoc comment documents a context pointer, but this is missing from the
> actual struct. Having some sort of driver_data (1) field here would be good IMHO.

Fine for me.

I left it out because nothing would be using it at the moment and
it would have been easy to add when needed.

Do you want me to send a v2 for that or would you add it when merging?
(Or remove the spurious comment)

> Regards,
> 
> Hans
> 
> 1) "context" is fine, but AFAIK almost all other foo_device_id structs call this
> driver_data, so that would be more consistent IMHO.

Thomas