[RFC PATCH] dm: add cells_count parameter in *_count_phandle_with_args

[Patrick DELAUNAY <patrick.delaunay@st.com>]

Hi Simon,

> From: Simon Glass <sjg@chromium.org>
> Sent: mardi 22 septembre 2020 20:49
> 
> On Thu, 10 Sep 2020 at 10:44, Patrick Delaunay <patrick.delaunay@st.com>
> wrote:
> >
> > The cell_count argument is required when cells_name is NULL.
> >
> > This patch adds this parameter in live tree API
> > - of_count_phandle_with_args
> > - ofnode_count_phandle_with_args
> > - dev_count_phandle_with_args
> >
> > This parameter solves issue when these API is used to count the number
> > of element of a cell without cell name. This parameter allow to force
> > the size cell.
> >
> > For example:
> >   count = dev_count_phandle_with_args(dev, "array", NULL, 3);
> >
> > Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
> > ---
> > I push today this RFC.
> >
> > It is linked to previous serie [1] but it is not a blocking point
> > today as no user use this API with cells_name = NULL
> > + dev_count_phandle_with_args
> > + ofnode_count_phandle_with_args
> >
> > But I think it is the good time to modify these functions as they are
> > not hugely used: it is the proposition in this RFC.
> >
> > It is just a RFC because I don't sure if I can modify the existing API
> > even if parameters are aligned with *_parse_phandle_with_args.
> >
> > I can also to add new APIs to use when cells_name is NULL:
> > + dev_count_phandle_with_cell_count(node, list_name, cell_count);
> > + ofnode_count_phandle_with_cell_count(node, list_name, cell_count);
> >
> > and raise a error if cells_name == NULL in existing function
> > + dev_count_phandle_with_args
> > + ofnode_count_phandle_with_args
> >
> > [1] http://patchwork.ozlabs.org/project/uboot/list/?series=200899
> >     "dm: add cells_count parameter in live DT APIs of_parse_phandle_with_args"
> >
> >
> >  board/st/stm32mp1/stm32mp1.c    | 2 +-
> >  drivers/clk/clk-uclass.c        | 4 ++--
> >  drivers/core/of_access.c        | 7 ++++---
> >  drivers/core/ofnode.c           | 6 +++---
> >  drivers/core/read.c             | 5 +++--
> >  drivers/phy/phy-uclass.c        | 2 +-
> >  drivers/reset/reset-uclass.c    | 2 +-
> >  drivers/usb/host/ehci-generic.c | 4 ++--
> >  include/dm/of_access.h          | 4 +++-
> >  include/dm/ofnode.h             | 3 ++-
> >  include/dm/read.h               | 8 +++++---
> >  11 files changed, 27 insertions(+), 20 deletions(-)'
> 
> Reviewed-by: Simon Glass <sjg@chromium.org>
> 
> A test would go a long way here.

Sure, I will add a test in the real patch...

I send RFC without test just to be sure that adding parameter in  *_count_phandle_with_args()
is a better solution than adding a new API.

Proposition 1 (it is the RFC content): add argument in current API

Example:

of_count_phandle_with_args(node, "clocks", "#clock-cells", 0);
ofnode_count_phandle_with_args(node, "resets", "#reset-cells", 0);
dev_count_phandle_with_args(node, "phys", "#phy-cells", 0);

dev_count_phandle_with_args(node, "test", NULL, 3);
ofnode_count_phandle_with_args(node, "test", NULL, 3);


Proposition 2: new API *count_phandle_with_cell_count

of_count_phandle_with_args(node, "clocks", "#clock-cells");
ofnode_count_phandle_with_args(node, "resets", "#reset-cells");
dev_count_phandle_with_args(node, "phys", "#phy-cells");

dev_count_phandle_with_fixed_args(node, "test", 3);
ofnode_count_phandle_with_fixed_args(node, "test", 3);

I think that Proposition 1 (this RFC) is more clear because parameters are aligned
with other API *read_phandle_with_args

But proposition 2 is align with Linux API
- of_count_phandle_with_args
- of_parse_phandle_with_fixed_args
And avoid to change all the current users of *count_phandle_with_args

Patrick