On Sun, 2022-10-30 at 20:18 -0700, Guenter Roeck wrote: > On Sun, Oct 30, 2022 at 01:06:45PM +1100, Frank Crawford wrote: > > ... > > I can see that you are for whatever reason completely opposed to > using > the dmi callback function. Fine, but then don't introduce another one > just to have a callback function and in case it may possibly be > needed > sometime in the future. Introduce it if and when it is needed, and > only > then. The same applies to all other infrastructure: Introduce it if > and > only if it is needed for more than one use case, not because it may > be > needed for some unspecified use case some time in the future. No, you get me wrong, I'm not completely opposed, although partly it was a misunderstanding on my part, as there are no good examples of the use of callbacks for DMI tables in the codes I looked though. I had to go looking into the actual functions to see how they are supposed to be used. Also, the most of the items I raise are not future cases for me, as they are in the out-of-tree version, and I am working on keeping that roughly in sync with the in-tree version, while I work on trying to merge most of that functionality in. Anyway, you should see that I've now submitted a v3 patch, which I believe does use the callback correctly, and is generally suitable for handling future updates as well. > > Thanks, > Guenter Regards Frank
