On Tue, Feb 07, 2023 at 09:45:46AM +1300, Daniel Beer wrote: > We did discuss this a while back when the driver first went in. > Unfortunately the vendor software tools provide configuration for the > part in the form of a sequence of raw register writes, including > explicit page changes: > https://lore.kernel.org/lkml/Yd85bjKEX9JnoOlI@xxxxxxxxxxxxx/ That seems surmountable, either bypassing regmap or parsing the configuration files. > Aside from this, I have two other practical issues. > The first is that I'm not sure how exactly to implement the paging > scheme in terms of regmap_range_cfg (assuming this is what you're > referring to). This chip has multi-level paging (books/pages), with the > book selection register itself requiring paging to access. A sequence of That's absolutely fine, this isn't the first device which has such a setup and the code handles nested windows fine. > Secondly, the patches as submitted here have been tested, but I don't > currently have access to hardware. I'm very hesitant to make a > significant change without retesting and leave the driver in a broken > state again. Presumably someone does given that the problem was noticed?
Attachment:
signature.asc
Description: PGP signature