Re: [PATCH] mod_devicetable.h: Split into per subsystem headers

["Uwe Kleine-König (The Capable Hub)" <u.kleine-koenig@baylibre.com>]

[-- Attachment #1: Type: text/plain, Size: 2155 bytes --]

Hello,

On Tue, Jun 23, 2026 at 06:13:29PM +0200, Uwe Kleine-König (The Capable Hub) wrote:
> On Tue, Jun 23, 2026 at 07:54:12AM -0700, Linus Torvalds wrote:
> > On Tue, 23 Jun 2026 at 07:47, Uwe Kleine-König (The Capable Hub)
> > <u.kleine-koenig@baylibre.com> wrote:
> > >
> > > So in the end I expect a handful of driver patches plus switching the
> > > the most important (or maybe all) <linux/$something.h>s from
> > > <linux/mod_devicetable.h> to <linux/mod_device_id/$something>.
> > >
> > > I guess that is ok then, too?
> > 
> > Sure, that sounds fine. I was expecting lots of files to include the
> > mod_device_id/xyz.h files by hand, since that's what currently happens
> > with that mod_devicetable.h file:
> > 
> >   $ git grep linux/mod_devicetable.h  | wc -l
> >   1621
> 
> These are mostly .c files. Currently I'm concentrating on the header
> files only:
> 
> $ find -name \*.h | xargs grep linux/mod_devicetable.h | wc -l
> 91
> 
> For these the effort-impact ratio is much better. Agreed that for .c
> files your script approach sounds right. Up to now I wasn't aware that
> there are that many .c files including <linux/mod_devicetable.h>
> directly.
> 
> Assuming I create such a script, what would be the approach to apply the
> resulting patch? Do you want to run and commit after -rc1? The
> alternative would be to submit a patch per subsystem.

Heads up: It's not as easy as anticipated. Two (similar) types of
difficulties I hit are:

 - drivers/platform/x86/msi-ec.c doesn't include
   <linux/mod_devicetable.h> but (e.g.) <linux/acpi.h>. The latter used
   to provide a definition for dmi_device_id. So when replacing
   <linux/mod_devicetable.h> in <linux/acpi.h> by
   <linux/mod_device_id/acpi.h>, drivers/platform/x86/msi-ec.c fails to
   compile due to missing dmi_device_id.

 - <linux/of_platform.h> makes use of resource_size_t. That was defined
   before via <linux/mod_devicetable.h> -> <linux/types.h>, but
   <linux/mod_device_id/of.h> doesn't include <linux/types.h>. Just
   adding that to be sure not to introduce build regessions is sad.

Best regards
Uwe

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]