Re: [PATCH v2 10/15] nvmem: core: add nvmem-layout support

[Marco Felsch <m.felsch@pengutronix.de>]

On 25-08-20, Sascha Hauer wrote:
> On Mon, Aug 18, 2025 at 07:44:44PM +0200, Marco Felsch wrote:
> > This ports the Linux nvmem-layout core infrastructure required for the
> > actual nvmem-layout drivers. This commit doesn't add any specific
> > nvmem-layout driver.
> > 
> > A nvmem-layout is the new way to describe the nvmem storage format. The
> > nvmem-layout driver parses the nvmem sotrage format and provides the
> > information via nvmem-cells.
> > 
> > Signed-off-by: Marco Felsch <m.felsch@pengutronix.de>
> > ---
> >  drivers/nvmem/Kconfig          |   7 ++
> >  drivers/nvmem/Makefile         |   3 +
> >  drivers/nvmem/core.c           | 100 +++++++++++++++++++++++-
> >  drivers/nvmem/internals.h      |  22 ++++++
> >  drivers/nvmem/layouts.c        | 173 +++++++++++++++++++++++++++++++++++++++++
> >  drivers/nvmem/layouts/Kconfig  |  13 ++++
> >  drivers/nvmem/layouts/Makefile |   7 ++
> >  include/linux/nvmem-provider.h |  63 +++++++++++++++
> >  8 files changed, 385 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/nvmem/Kconfig b/drivers/nvmem/Kconfig
> > +struct device_node *of_nvmem_layout_get_container(struct nvmem_device *nvmem)
> > +{
> > +	return of_get_child_by_name(nvmem->dev.of_node, "nvmem-layout");
> 
> This also needs an additional
> 
> 	if (!nvmem->dev.of_node)
> 		return NULL;
> 
> check. This also goes down to of_get_child_by_name doing a
> for_each_child_of_node() which is defines as:
> 
> #define for_each_child_of_node(parent, child) \
> 	list_for_each_entry(child, &parent->children, parent_list)
> 
> list_for_each_entry() does not work with NULL pointers. We might want to
> change this for Linux compatibility.

Thanks for testing and spotting this. I ran tests but of course all my
nvmem devices had an of_node :(

Regards,
  Marco