Hi Niklas, On 09/08/2019 13:23, Niklas Söderlund wrote: > Hi Kieran, > > On 2019-08-09 13:12:49 +0100, Kieran Bingham wrote: >> On 09/08/2019 13:04, Niklas Söderlund wrote: >>> Hi Kieran, >>> >>> Thanks for your feedback. >>> >>> On 2019-08-09 09:09:43 +0100, Kieran Bingham wrote: >>>> Hi Niklas, >>>> >>>> This should be at least v5. >>> >>> I don't agree ;-) This is a "new" series where multiple streams are not >>> supported and there are no external dependencies. To indicate this I >> >> I'm afraid there's nothing new about a version of this series with >> support for only a single stream. >> >> See version 2 of my series: >> >> https://lore.kernel.org/linux-media/20180808165559.29957-4-kieran.bingham@xxxxxxxxxxxxxxxx/ > > Was not aware of this posting, then yes I do agree with you this should > have been v5. Sorry about that. > >> >> >>> reset the version. I don't feel strongly about this next submission can >>> remedy this if you do feel strongly about it. >> >> I find it very confusing to have reset the version but kept exactly the >> same patch title. >> >> It's a fork of the series :D > > :-) > > I think this proves we need to get this driver upstream so we can start > submitting patches towards something and not brew our our brand of stuff > we have laying around and find all over the place. I agree, It's unfortunate that the direction has rotated too many times for this driver. Originally it was posted with VC support as that was our development, then I removed the VC support (v2, v3) to aim to get it upstreamed. I can't recall now why V4 occurred with VC brought back in - but perhaps it was because the VC work was in active development at the time - and I thought it was making progress, but it soon became a blocker - and was not worked on until Jacopo started looking at it again recently - and became a complete block to reposting any V5 of MAX9286 - as the VC work was out of date and not being rebased. Anyway, I agree - that getting a non-multi-stream capable version of this integrated is very beneficial, so please do continue and post a v5 (or v6 if you deem more appropriate). Please make sure all pending review comments are explored (especially the one from Rob Herring about using an i2c-mux node) before posting your next version though. -- Kieran >>>> Did you take the last v4 and work from there? >>>> I have made changes since the last posting. Did you get an update from >>>> my branches? >>>> >>>> What changes have you made to this posting compared to whichever >>>> patch-base you have taken to start from? >>> >>> I took my latest known good state and diffed it with all gmsl branches i >>> could find picked what seamed most recent. Then I removed multiplexed >>> stream support, fixed a few todos in error paths to clean up notifiers >>> and unified naming of the private data structure. >> >> >> I'll diff your version with my latest. >> >> There were other review comments worked on from Sakari too., and there >> is an outstanding comment from Rob to complete on the bindings from v4. >> >> -- >> Kieran >> >> >> >