On Tue, 15 Jan 2019 11:09:51 +0800 Daniel Drake <drake@xxxxxxxxxxxx> wrote: > Hi Denis, > > On Tue, Jan 15, 2019 at 8:18 AM Denis CIOCCA <denis.ciocca@xxxxxx> wrote: > > I am not familiar with ACPI and I am not sure why axis orientation is not following the standard there. > > What I wonder is if this is related to SMO8840, Acer or in general to ST accelerometers integration. Your patch seems to be fine technically but are you able to confirm it will works even outside the 3 parts tested? The other accel supported by ACPI probing SMO8A90 is affected also? > > The _ONT data format was explained to us by a vendor support contact > at ST. (however, the fact that the _ONT transformation must be applied > after first applying another translation was determined experimentally > by me). > > We have seen this _ONT data present in ACPI tables from multiple > product vendors, on SMO8820, SMO8821 and SMO8840 devices, so I believe > there is no doubt that it is a ST-specific thing. > > The specific transformation behaviour in my patch has only been tested > on products with SMO8840 though, since those are the only ones we have > on hand. > > Daniel I suspect we aren't going to come to a firm conclusion on this any time soon. As it only effects devices that do provide _ONT, it is fairly unlikely to cause a regression. Ideally we would know if it worked on the other devices where this is known to exist but such is life. If people agree that I'm right in patch 1 not being necessary and we don't have any objections to merging this I'll do so, probably next weekend now. Thanks, Jonathan
