v2022.08.1

Full Changelog: v2022.08...v2022.08.1

v2021.02 Release

v2021.02 Release:

- Stricter root node and simple-bus unit-address checks

- Better undocumented compatible string checking

- Rework fixups yet again. Now all fixups are done in one spot (after
  check_schema()). Matching of unit suffixes is added for integer
  fixups.

- Make assigned-clocks* properties always allowed.

- Add percentage as unit suffix

- Add meta-schemas for hwlock and nvmem bindings

v2020.12 Release

dtschema v2020.12:

- Add graph schema

- Add IIO schemas

- Allow draft2019.09 $def keyword

- Fix warning filtering on disabled child nodes

- Drop 'allOf' for $ref (draft2019.09 syntax)

v2020.08.1 Release

dt-validate: Print the schema file causing the error

Since multiple schemas can be applied to a node, it's not always clear
where the source of an error was. The jsonschema errors also don't have
any way to get back to the schema file, but we have at least the starting
schema file, so let's print it.

If the schema file references other files, then those files could be the
actual source. There's not currently a way to track $refs, so we only get
the starting point.

Signed-off-by: Rob Herring <robh@kernel.org>

v2020.05 Release

meta-schemas: Ensure json-schema vocab aren't DT property names

It's very easy to define json-schema vocabulary as DT properties by
mistake if the indentation is wrong.

There's unlikely to be much overlap. 'type' is one such case though.

Signed-off-by: Rob Herring <robh@kernel.org>

v2019.11 Release

dtschema: Rework int array to matrix fixups

Using _is_int_schema to match int arrays doesn't work if we have a list
of 'items' with just descriptions as that looks the same as a string or
phandle list. To fix this, let's also look for a $ref to a 'uint*-array'
type in addition to integer keywords. As part of this, _fixup_int_array
becomes redundant and can be removed.

The items size fixup needs to be done after all the other fixups and after
all the recursion. Otherwise, we miss some instances.

Signed-off-by: Rob Herring <robh@kernel.org>