On Thu, Mar 12, 2020 at 12:05:41PM -0700, Eric Biggers wrote: > Of course, if someone actually posts patches to support hardware that diverges > from the UFS standard in new and "exciting" ways (whether it's another vendor's > hardware or future Qualcomm hardware) then they'll need to post any variant > operation(s) they need. They need to be targetted to only the specific quirk(s) > needed, so that drivers don't have to unnecessarily re-implement stuff. I think the only sane answer is that we only support hardware upstream that conforms to the standard and we skip all that bullshit. The whole point of the standard is that things just work, and we should not give vendors a wild card to come up with bullshit like the interfaces handled in this patchset.
