On Fri, Feb 14, 2025 at 07:25:59AM -0800, Doug Anderson wrote: > Hi, > > On Fri, Feb 14, 2025 at 12:59 AM Jerome Brunet <jbrunet@xxxxxxxxxxxx> wrote: > > > > > One think that's always felt a bit meh to me is this id number stuff, > > > I just threw in 666 for meme value. > > > > :) > > > > > The whole thing seems super > > > arbitrary, do any of the users of this helper actually put meaningful > > > values into the id parameter? > > > > In example changes I've sent, no. > > > > In other simple usages (those using container_of()) of the auxiliary > > bus, I think there are a few that uses 0 and 1 for 2 instances. > > > > I guess your question is "do we really need this parameter here ?" > > > > We could remove it and still address 90% of the original target. > > > > Keeping it leaves the door open in case the figure above does not hold > > and it is pretty cheap to do. It could also enable drivers requiring an > > IDA to use the helper, possibly. > > FWIW, once you resolve the conflicts in drm-misc with ti-sn65dsi86 > you'll need the ID value. ;-) > > There was a big-ol' discussion here: > > https://lore.kernel.org/r/8c2df6a903f87d4932586b25f1d3bd548fe8e6d1.1729180470.git.geert+renesas@xxxxxxxxx > > I eventually pushed v2 of the patch: > > https://lore.kernel.org/r/7a68a0e3f927e26edca6040067fb653eb06efb79.1733840089.git.geert+renesas@xxxxxxxxx Think it makes sense to have a "simplified" wrapper for the cases where the id has no meaning then? Not really a fan of the drivers coming up with magic numbers.
Attachment:
signature.asc
Description: PGP signature
