Add optional '#nvmem-cell-cells' #89
Conversation
Add support for having arguments to a phandle to a nvmem-cell. To be backwards compatible, we have to make the '#nvmem-cell-cells' property optional. If it's missing no arguments are expected. Signed-off-by: Michael Walle <michael@walle.cc>
'nvmem-cells' can now have a '#nvmem-cell-cells' property. Handle that case correctly. If it is missing, no arguments are expected. _get_phandle_arg_size() already handles both cases correctly. Signed-off-by: Michael Walle <michael@walle.cc>
mwalle
force-pushed
the
feature-nvmem-cells-args
branch
from
eef6d0d
to
77dd9ac
Compare
|
I wasn't aware of this work and meanwhile sent a patch, see: Hopefully we can get that |
| $ref: "cell.yaml#/array" | ||
| oneOf: | ||
| - $ref: "cell.yaml#/array" | ||
| - $ref: "/schemas/types.yaml#/definitions/phandle-array" |
There was a problem hiding this comment.
meta-schemas are what check the schemas. What you are adding is what belongs in the schema, and is already there (in the kernel tree).
| @@ -12,6 +12,8 @@ properties: | |||
| nvmem-cells: | |||
| $ref: "cell.yaml#/array" | |||
|
|
|||
| '#nvmem-cell-cells': true | |||
|
|
|||
There was a problem hiding this comment.
Not needed in meta-schema. What you do need is:
#nvmem-cell-cells:
enum: [ 0, 1 ]
In nvmem.yaml in the kernel. (Really, I'd like to move those here.)
The exact constraints depends on if you want to globally define how many and what the cells contain.
There was a problem hiding this comment.
Any plans for working on this? I guess after almost 2 years not really.
|
@robherring: I'm confused by your mention & suggestion to move that stuff into kernel. I believe (unless I got wrong impression) that you actually suggested adding
Does my |
|
@rmilecki My comment was the dtb.py change is what is needed here. I would like to see nvmem.yaml and nvmem-consumer.yaml moved from the kernel to dtschema, but that's a separate issue. For now, it's fine to add #nvmem-cell-cells to kernel's nvmem.yaml. |
|
We now have Does this pull request still make any sense? Or should it be closed? |
|
I believe at least the dtb.py part is needed still. |
| @@ -262,7 +262,6 @@ def fdt_scan_node(fdt, nodename, offset): | |||
| 'msi-ranges': '#interrupt-cells', | |||
| 'gpio-ranges': 3, | |||
|
|
|||
| 'nvmem-cells': None, | |||
There was a problem hiding this comment.
Independently, Rafał Miłecki reported a bug regarding nvmem-cells and I narrowed it to this line. Later I found this pull request.
I think this part is needed along with Rob's comment for last hunk...
|
I'll try to sort it out. |
Linux kernel NVMEM consumer bindings define phandle to NVMEM cells
("nvmem-cells"), thus we also want the common definition of property
defining number of cells encoding that specifier, so the
Suggested-by: Rob Herring <robh@kernel.org>
Reported-by: Michael Walle <michael@walle.cc>
Closes: devicetree-org/dt-schema#89
Reported-by: Rafał Miłecki <zajec5@gmail.com>
Closes: https://lore.kernel.org/linux-arm-kernel/20221121105830.7411-1-zajec5@gmail.com/#r
Closes: https://lore.kernel.org/all/bdf7751b-0421-485d-8382-26c084f09d7d@gmail.com/
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Linux kernel NVMEM consumer bindings define phandle to NVMEM cells
("nvmem-cells"), thus we also want the common definition of property
defining number of cells encoding that specifier, so the
Suggested-by: Rob Herring <robh@kernel.org>
Reported-by: Michael Walle <michael@walle.cc>
Closes: devicetree-org/dt-schema#89
Reported-by: Rafał Miłecki <zajec5@gmail.com>
Closes: https://lore.kernel.org/linux-arm-kernel/20221121105830.7411-1-zajec5@gmail.com/#r
Closes: https://lore.kernel.org/all/bdf7751b-0421-485d-8382-26c084f09d7d@gmail.com/
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
|
Fixed in commit 0053468 ("dtb: expect nvmem-cell-cells") |
Linux kernel NVMEM consumer bindings define phandle to NVMEM cells
("nvmem-cells"), thus we also want the common definition of property
defining number of cells encoding that specifier, so the
Suggested-by: Rob Herring <robh@kernel.org>
Reported-by: Michael Walle <michael@walle.cc>
Closes: devicetree-org/dt-schema#89
Reported-by: Rafał Miłecki <zajec5@gmail.com>
Closes: https://lore.kernel.org/linux-arm-kernel/20221121105830.7411-1-zajec5@gmail.com/#r
Closes: https://lore.kernel.org/all/bdf7751b-0421-485d-8382-26c084f09d7d@gmail.com/
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Linux kernel NVMEM consumer bindings define phandle to NVMEM cells
("nvmem-cells"), thus we also want the common definition of property
defining number of cells encoding that specifier, so the
Suggested-by: Rob Herring <robh@kernel.org>
Reported-by: Michael Walle <michael@walle.cc>
Closes: devicetree-org/dt-schema#89
Reported-by: Rafał Miłecki <zajec5@gmail.com>
Closes: https://lore.kernel.org/linux-arm-kernel/20221121105830.7411-1-zajec5@gmail.com/#r
Closes: https://lore.kernel.org/all/bdf7751b-0421-485d-8382-26c084f09d7d@gmail.com/
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Linux kernel NVMEM consumer bindings define phandle to NVMEM cells
("nvmem-cells"), thus we also want the common definition of property
defining number of cells encoding that specifier, so the
Suggested-by: Rob Herring <robh@kernel.org>
Reported-by: Michael Walle <michael@walle.cc>
Closes: devicetree-org/dt-schema#89
Reported-by: Rafał Miłecki <zajec5@gmail.com>
Closes: https://lore.kernel.org/linux-arm-kernel/20221121105830.7411-1-zajec5@gmail.com/#r
Closes: https://lore.kernel.org/all/bdf7751b-0421-485d-8382-26c084f09d7d@gmail.com/
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Linux kernel NVMEM consumer bindings define phandle to NVMEM cells
("nvmem-cells"), thus we also want the common definition of property
defining number of cells encoding that specifier, so the
Suggested-by: Rob Herring <robh@kernel.org>
Reported-by: Michael Walle <michael@walle.cc>
Closes: devicetree-org/dt-schema#89
Reported-by: Rafał Miłecki <zajec5@gmail.com>
Closes: https://lore.kernel.org/linux-arm-kernel/20221121105830.7411-1-zajec5@gmail.com/#r
Closes: https://lore.kernel.org/all/bdf7751b-0421-485d-8382-26c084f09d7d@gmail.com/
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20240224114516.86365-5-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Linux kernel NVMEM consumer bindings define phandle to NVMEM cells
("nvmem-cells"), thus we also want the common definition of property
defining number of cells encoding that specifier, so the
Suggested-by: Rob Herring <robh@kernel.org>
Reported-by: Michael Walle <michael@walle.cc>
Closes: devicetree-org/dt-schema#89
Reported-by: Rafał Miłecki <zajec5@gmail.com>
Closes: https://lore.kernel.org/linux-arm-kernel/20221121105830.7411-1-zajec5@gmail.com/#r
Closes: https://lore.kernel.org/all/bdf7751b-0421-485d-8382-26c084f09d7d@gmail.com/
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20240224114516.86365-5-srinivas.kandagatla@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ upstream commit: d28c853 ]
As requested in https://lore.kernel.org/linux-devicetree/20220912192038.GA1661550-robh@kernel.org/, add the property in dtschema.
The second patch is my try to also add support to the dtschema tooling. I'm not sure, if the entry for the 'nvmem-cells' in phandle_args ever worked, because the type of nvmem-cells wasn't a phandle-array. Thus fixup_phandles() seems to skip that property.