mbox series

[v2,0/9] soc: remove direct accesses to of_root from drivers/soc/

Message ID 20260223-soc-of-root-v2-0-b45da45903c8@oss.qualcomm.com (mailing list archive)
Headers
Series soc: remove direct accesses to of_root from drivers/soc/ |

Message

Bartosz Golaszewski Feb. 23, 2026, 1:37 p.m. UTC 
linux/of.h declares a set of variables providing addresses of certain
key OF nodes. The pointers being variables can't profit from stubs
provided for when CONFIG_OF is disabled which means that drivers
accessing these variables can't profit from CONFIG_COMPILE_TEST=y
coverage.

There are drivers under drivers/soc/ that access the of_root node. This
series introduces new OF helpers for reading the machine compatible and
model strings, exports an existing SoC helper that reads the machine
string from the root node and finally replaces all direct accesses to
of_root with new or already existing helper functions.

Merging strategy: first two patches should be either acked by Rob or
picked up into an immutable branch based on v7.0-rc1, the rest can go
through the SoC tree.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
---
Changes in v2:
- rename of_machine_get_compatible() to of_machine_read_compatible() and
  add the index argument
- add a stub for of_machine_read_compatible() for !CONFIG_OF
- provide of_machine_read_model() to avoid having to look-up the root
  node via of_find_node_by_path("/")
- rename soc_device_get_machine() to soc_attr_read_machine() before
  making it an exported symbol
- Link to v1: https://lore.kernel.org/r/20260119-soc-of-root-v1-0-32a0fa9a78b4@oss.qualcomm.com

---
Bartosz Golaszewski (9):
      of: provide of_machine_read_compatible()
      of: provide of_machine_read_model()
      base: soc: order includes alphabetically
      base: soc: rename and export soc_device_get_machine()
      soc: fsl: guts: don't access of_root directly
      soc: imx8m: don't access of_root directly
      soc: imx9: don't access of_root directly
      soc: renesas: don't access of_root directly
      soc: sunxi: mbus: don't access of_root directly

 drivers/base/soc.c                | 23 ++++++++++-------------
 drivers/of/base.c                 | 28 ++++++++++++++++++++++++++++
 drivers/soc/fsl/guts.c            | 12 +++---------
 drivers/soc/imx/soc-imx8m.c       | 11 +++--------
 drivers/soc/imx/soc-imx9.c        |  4 ++--
 drivers/soc/renesas/renesas-soc.c |  7 ++++++-
 drivers/soc/sunxi/sunxi_mbus.c    |  2 +-
 include/linux/of.h                | 14 ++++++++++++++
 include/linux/sys_soc.h           | 10 ++++++++++
 9 files changed, 77 insertions(+), 34 deletions(-)
---
base-commit: 6de23f81a5e08be8fbf5e8d7e9febc72a5b5f27f
change-id: 20260119-soc-of-root-77c86c54010f

Best regards,

Comments

Rob Herring (Arm) Feb. 24, 2026, 6:38 p.m. UTC | #1 
On Mon, Feb 23, 2026 at 02:37:15PM +0100, Bartosz Golaszewski wrote:
> linux/of.h declares a set of variables providing addresses of certain
> key OF nodes. The pointers being variables can't profit from stubs
> provided for when CONFIG_OF is disabled which means that drivers
> accessing these variables can't profit from CONFIG_COMPILE_TEST=y
> coverage.
> 
> There are drivers under drivers/soc/ that access the of_root node. This
> series introduces new OF helpers for reading the machine compatible and
> model strings, exports an existing SoC helper that reads the machine
> string from the root node and finally replaces all direct accesses to
> of_root with new or already existing helper functions.
> 
> Merging strategy: first two patches should be either acked by Rob or
> picked up into an immutable branch based on v7.0-rc1, the rest can go
> through the SoC tree.

SoC tree is good.

> 
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
> ---
> Changes in v2:
> - rename of_machine_get_compatible() to of_machine_read_compatible() and
>   add the index argument
> - add a stub for of_machine_read_compatible() for !CONFIG_OF
> - provide of_machine_read_model() to avoid having to look-up the root
>   node via of_find_node_by_path("/")
> - rename soc_device_get_machine() to soc_attr_read_machine() before
>   making it an exported symbol
> - Link to v1: https://lore.kernel.org/r/20260119-soc-of-root-v1-0-32a0fa9a78b4@oss.qualcomm.com
> 
> ---
> Bartosz Golaszewski (9):
>       of: provide of_machine_read_compatible()
>       of: provide of_machine_read_model()
>       base: soc: order includes alphabetically
>       base: soc: rename and export soc_device_get_machine()
>       soc: fsl: guts: don't access of_root directly
>       soc: imx8m: don't access of_root directly
>       soc: imx9: don't access of_root directly
>       soc: renesas: don't access of_root directly
>       soc: sunxi: mbus: don't access of_root directly

For all but patch 8,

Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
Bartosz Golaszewski Feb. 26, 2026, 9:28 a.m. UTC | #2 
On Tue, Feb 24, 2026 at 7:38 PM Rob Herring <robh@kernel.org> wrote:
>
> On Mon, Feb 23, 2026 at 02:37:15PM +0100, Bartosz Golaszewski wrote:
> > linux/of.h declares a set of variables providing addresses of certain
> > key OF nodes. The pointers being variables can't profit from stubs
> > provided for when CONFIG_OF is disabled which means that drivers
> > accessing these variables can't profit from CONFIG_COMPILE_TEST=y
> > coverage.
> >
> > There are drivers under drivers/soc/ that access the of_root node. This
> > series introduces new OF helpers for reading the machine compatible and
> > model strings, exports an existing SoC helper that reads the machine
> > string from the root node and finally replaces all direct accesses to
> > of_root with new or already existing helper functions.
> >
> > Merging strategy: first two patches should be either acked by Rob or
> > picked up into an immutable branch based on v7.0-rc1, the rest can go
> > through the SoC tree.
>
> SoC tree is good.
>
> For all but patch 8,
>
> Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
>

FYI Rob also reviewed patch 8 now. Who would pick the patches up? Greg
KH? There does not seem to be a centralized SoC maintainer in
MAINTAINERS?

Bart
Greg KH March 12, 2026, 3:18 p.m. UTC | #3 
On Thu, Feb 26, 2026 at 10:28:26AM +0100, Bartosz Golaszewski wrote:
> On Tue, Feb 24, 2026 at 7:38 PM Rob Herring <robh@kernel.org> wrote:
> >
> > On Mon, Feb 23, 2026 at 02:37:15PM +0100, Bartosz Golaszewski wrote:
> > > linux/of.h declares a set of variables providing addresses of certain
> > > key OF nodes. The pointers being variables can't profit from stubs
> > > provided for when CONFIG_OF is disabled which means that drivers
> > > accessing these variables can't profit from CONFIG_COMPILE_TEST=y
> > > coverage.
> > >
> > > There are drivers under drivers/soc/ that access the of_root node. This
> > > series introduces new OF helpers for reading the machine compatible and
> > > model strings, exports an existing SoC helper that reads the machine
> > > string from the root node and finally replaces all direct accesses to
> > > of_root with new or already existing helper functions.
> > >
> > > Merging strategy: first two patches should be either acked by Rob or
> > > picked up into an immutable branch based on v7.0-rc1, the rest can go
> > > through the SoC tree.
> >
> > SoC tree is good.
> >
> > For all but patch 8,
> >
> > Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
> >
> 
> FYI Rob also reviewed patch 8 now. Who would pick the patches up? Greg
> KH? There does not seem to be a centralized SoC maintainer in
> MAINTAINERS?

Sure, let me take them now, thanks.

greg k-h
Bartosz Golaszewski March 13, 2026, 8:36 a.m. UTC | #4 
On Thu, 12 Mar 2026 16:18:38 +0100, Greg Kroah-Hartman
<gregkh@linuxfoundation.org> said:
> On Thu, Feb 26, 2026 at 10:28:26AM +0100, Bartosz Golaszewski wrote:
>> On Tue, Feb 24, 2026 at 7:38 PM Rob Herring <robh@kernel.org> wrote:
>> >
>> > On Mon, Feb 23, 2026 at 02:37:15PM +0100, Bartosz Golaszewski wrote:
>> > > linux/of.h declares a set of variables providing addresses of certain
>> > > key OF nodes. The pointers being variables can't profit from stubs
>> > > provided for when CONFIG_OF is disabled which means that drivers
>> > > accessing these variables can't profit from CONFIG_COMPILE_TEST=y
>> > > coverage.
>> > >
>> > > There are drivers under drivers/soc/ that access the of_root node. This
>> > > series introduces new OF helpers for reading the machine compatible and
>> > > model strings, exports an existing SoC helper that reads the machine
>> > > string from the root node and finally replaces all direct accesses to
>> > > of_root with new or already existing helper functions.
>> > >
>> > > Merging strategy: first two patches should be either acked by Rob or
>> > > picked up into an immutable branch based on v7.0-rc1, the rest can go
>> > > through the SoC tree.
>> >
>> > SoC tree is good.
>> >
>> > For all but patch 8,
>> >
>> > Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
>> >
>>
>> FYI Rob also reviewed patch 8 now. Who would pick the patches up? Greg
>> KH? There does not seem to be a centralized SoC maintainer in
>> MAINTAINERS?
>
> Sure, let me take them now, thanks.
>
> greg k-h
>

In the meantime Geert sent an alternative to patch 8/9 so this single one can
be dropped from the series.

Thanks,
Bartosz
Greg KH March 13, 2026, 9:22 a.m. UTC | #5 
On Fri, Mar 13, 2026 at 01:36:49AM -0700, Bartosz Golaszewski wrote:
> On Thu, 12 Mar 2026 16:18:38 +0100, Greg Kroah-Hartman
> <gregkh@linuxfoundation.org> said:
> > On Thu, Feb 26, 2026 at 10:28:26AM +0100, Bartosz Golaszewski wrote:
> >> On Tue, Feb 24, 2026 at 7:38 PM Rob Herring <robh@kernel.org> wrote:
> >> >
> >> > On Mon, Feb 23, 2026 at 02:37:15PM +0100, Bartosz Golaszewski wrote:
> >> > > linux/of.h declares a set of variables providing addresses of certain
> >> > > key OF nodes. The pointers being variables can't profit from stubs
> >> > > provided for when CONFIG_OF is disabled which means that drivers
> >> > > accessing these variables can't profit from CONFIG_COMPILE_TEST=y
> >> > > coverage.
> >> > >
> >> > > There are drivers under drivers/soc/ that access the of_root node. This
> >> > > series introduces new OF helpers for reading the machine compatible and
> >> > > model strings, exports an existing SoC helper that reads the machine
> >> > > string from the root node and finally replaces all direct accesses to
> >> > > of_root with new or already existing helper functions.
> >> > >
> >> > > Merging strategy: first two patches should be either acked by Rob or
> >> > > picked up into an immutable branch based on v7.0-rc1, the rest can go
> >> > > through the SoC tree.
> >> >
> >> > SoC tree is good.
> >> >
> >> > For all but patch 8,
> >> >
> >> > Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
> >> >
> >>
> >> FYI Rob also reviewed patch 8 now. Who would pick the patches up? Greg
> >> KH? There does not seem to be a centralized SoC maintainer in
> >> MAINTAINERS?
> >
> > Sure, let me take them now, thanks.
> >
> > greg k-h
> >
> 
> In the meantime Geert sent an alternative to patch 8/9 so this single one can
> be dropped from the series.

Ick, ok, I've now dropped that one commit from the tree.

thanks,

greg k-h