On 09/07/2024 22:48, Rob Herring wrote: > On Tue, Jul 9, 2024 at 11:15 AM Rob Herring <robh@xxxxxxxxxx> wrote: >> >> On Tue, Jul 09, 2024 at 10:44:01AM +0200, Krzysztof Kozlowski wrote: >>> There is a device in the wild with non-updatable firmware coming with >>> ACPI tables with rejected "ltr,ltrf216a" compatible. Linux kernel still >>> supports this device via ACPI PRP0001, however the compatible was never >>> accepted to bindings. Lack of bindings causes checkpatch.pl warning >>> about undocumented compatible. >> >> Why do we care? For checkpatch.pl I really don't. That hack check I >> wrote makes any string in binding docs a documented compatible. I have a >> better check using the schema written, but that would make checkpatch >> dependent on dtschema tools. So maybe just time to drop this check from >> checkpatch as we have other ways to check and track this. People still use checkpatch - both to actually test patches before sending and also to fix random existing issues. >> >> However, I do care about 'make dt_compatible_check'. Besides these ACPI >> cases, there's a bunch of cases that we'll never have schemas for. Like >> everything from Sparc... Old PowerMac stuff... So I would like to >> 'document' them just to exclude from dt_compatible_check. So perhaps >> this should be generalized. Sure, I can rewrite it to more generic. > > Here's my list of what's really not documented. It's just a grep of > the bindings of each compatible found by 'make dt_compatible_check'. > Probably anything with SUNW, ibm, amcc, or mpc5 is never going to be > documented. > > There are some false positives such as cases documented like "fsl,<chip>-guts". I'll come with something, maybe incomplete but it could grow later. Best regards, Krzysztof