Messages in this thread | | | Date | Sun, 28 Aug 2022 16:42:07 +0200 | From | Michael Walle <> | Subject | Re: [RFC PATCH v1 12/14] nvmem: layouts: rewrite the u-boot-env driver as a NVMEM layout |
| |
Am 2022-08-28 16:04, schrieb Rafał Miłecki: > On 25.08.2022 23:44, Michael Walle wrote: >> Instead of hardcoding the underlying access method mtd_read() and >> duplicating all the error handling, rewrite the driver as a nvmem >> layout which just uses nvmem_device_read() and thus works with any >> NVMEM device. >> >> But because this is now not a device anymore, the compatible string >> will have to be changed so the device will still be probed: >> compatible = "u-boot,env"; >> to >> compatible = "u-boot,env", "nvmem-cells"; >> >> "nvmem-cells" will tell the mtd layer to register a nvmem_device(). >> "u-boot,env" will tell the NVMEM that it should apply the u-boot >> environment layout to the NVMEM device. > > That's fishy but maybe we can ignore backward compatibility at > point.
As mentioned in the cover letter, this is why this is an RFC. I didn't see any users in the device tree, nor can I see how this would have been used anyway, because you cannot find a cell by its device tree node, because none is registered. So maybe we can still change the compatible string.
-michael
> Still you need to update DT binding.
| |