hid-input.c && vendor defined usage pages

hid-input.c && vendor defined usage pages

[Aseda Aboagye]

Perhaps this is a naive question, but why does `hidinput_configure_usage()` in
`hid-input.c` configure vendor defined usage pages? In actuality, it seems like
it's mainly handling some HP specific usage pages.  Why aren't those in some
`hid-hp.c` file that's selected by a kernel config?

Curiously,

--
Aseda Aboagye

Re: hid-input.c && vendor defined usage pages

[Dmitry Torokhov]

Hi Aseda,

On Tue, Sep 24, 2024 at 07:32:30PM -0500, Aseda Aboagye wrote:
> Perhaps this is a naive question, but why does `hidinput_configure_usage()` in
> `hid-input.c` configure vendor defined usage pages? In actuality, it seems like
> it's mainly handling some HP specific usage pages.  Why aren't those in some
> `hid-hp.c` file that's selected by a kernel config?

It is simply a historic artefact. These usages were mapped way before we
had HID sub-drivers and accompanying infrastructure.

Thanks.

-- 
Dmitry

Re: hid-input.c && vendor defined usage pages

[Aseda Aboagye]

On Wed, Oct 02, 2024 at 08:00:59AM -0700, Dmitry Torokhov wrote:
> Hi Aseda,
> 
> On Tue, Sep 24, 2024 at 07:32:30PM -0500, Aseda Aboagye wrote:
> > Perhaps this is a naive question, but why does `hidinput_configure_usage()` in
> > `hid-input.c` configure vendor defined usage pages? In actuality, it seems like
> > it's mainly handling some HP specific usage pages.  Why aren't those in some
> > `hid-hp.c` file that's selected by a kernel config?
> 
> It is simply a historic artefact. These usages were mapped way before we
> had HID sub-drivers and accompanying infrastructure.

Do you reckon that we're stuck with this behaviour now since migrating them
out, will likely break users who don't have the appropriate sub-drivers
selected in their kernel config?

Thanks,

-- 
Aseda Aboagye

Re: hid-input.c && vendor defined usage pages

[Dmitry Torokhov]

On Mon, Oct 14, 2024 at 10:15:03PM +0000, Aseda Aboagye wrote:
> On Wed, Oct 02, 2024 at 08:00:59AM -0700, Dmitry Torokhov wrote:
> > Hi Aseda,
> > 
> > On Tue, Sep 24, 2024 at 07:32:30PM -0500, Aseda Aboagye wrote:
> > > Perhaps this is a naive question, but why does `hidinput_configure_usage()` in
> > > `hid-input.c` configure vendor defined usage pages? In actuality, it seems like
> > > it's mainly handling some HP specific usage pages.  Why aren't those in some
> > > `hid-hp.c` file that's selected by a kernel config?
> > 
> > It is simply a historic artefact. These usages were mapped way before we
> > had HID sub-drivers and accompanying infrastructure.
> 
> Do you reckon that we're stuck with this behaviour now since migrating them
> out, will likely break users who don't have the appropriate sub-drivers
> selected in their kernel config?

We typically do not commit to keeping forward or backward compatibility
for the kernel config options, it is distribution's task to configure
the kernel properly, so I think it can be changed if desired.

Thanks.

-- 
Dmitry