Hi Srinivas, srinivas.kandagatla@xxxxxxxxxx wrote on Fri, 10 Mar 2023 09:29:18 +0000: > On 10/03/2023 07:51, Rafał Miłecki wrote: > > From: Rafał Miłecki <rafal@xxxxxxxxxx> > > > > With support for "fixed-layout" binding we can use now "nvmem-layout" > > even for fixed NVMEM cells. Use that in the base example as it should be > > preferred over placing cells directly in the device node. > > > Fixed layouts are the core part of nvmem, am not sure why you want to deprecate this. Either we derive the cell information dt or via layouts or some post processing they should still endup as fixed layouts. > this way the core part is always same irrespective of where the cell info comes from. Actually I was suspicious as first but we had a very similar case in mtd: - People need partitioning so we add partitions in the mtd node - People need more advanced partitioning and partitioning becomes a mess so we containerize everything in a "partitions" subnode. It definitely improves the readability and makes the code more resilient: outside of the container, it's not a partition, period. I believe that's what Rafał is trying to anticipate. Just moving the fixed cells declaration into a container (and we have one already: nvmem-layout). Thanks, Miquèl
