Re: [PATCH 1/3] iio: accel: bmc150: Improve ACPI enumeration support

[Jonathan Cameron <jic23@kernel.org>]

On Sat, 28 Nov 2020 14:32:49 +0100
Hans de Goede <hdegoede@redhat.com> wrote:

> Hi,
> 
> On 11/28/20 2:16 PM, Jonathan Cameron wrote:
> > On Wed, 25 Nov 2020 11:49:15 +0100
> > Hans de Goede <hdegoede@redhat.com> wrote:
> >   
> >> Hi,
> >>
> >> On 11/25/20 11:41 AM, Andy Shevchenko wrote:  
> >>> On Wed, Nov 25, 2020 at 10:37 AM Hans de Goede <hdegoede@redhat.com> wrote:    
> >>>>
> >>>> Hi All,
> >>>>
> >>>> This series improves support for ACPI enumeration of bmc150 accels
> >>>> described by an ACPI node with an ACPI hw-id of BOSC0200:
> >>>>
> >>>> 1. Add support for nodes which describe 2 acceleromers in a single node,    
> >>>
> >>> accelerometers
> >>>     
> >>>> fixing: https://bugzilla.kernel.org/show_bug.cgi?id=198671
> >>>>
> >>>> 2. Add support for reading the mount-matrix from the ACPI node.
> >>>>
> >>>> This is done in patches 2 - 3, patch 1 is a trivial cleanup which I noticed
> >>>> could be done while working on this.
> >>>>
> >>>> Patch 2 is based on an earlier patch for this from Jeremy Cline:
> >>>> https://lore.kernel.org/r/010001602cf53153-39ad69f1-1b39-4e6d-a748-9455a16c2fbd-000000@email.amazonses.com
> >>>>
> >>>> That patch was put on hold because normally ACPI nodes which describe
> >>>> multiple i2c-clients in a single node are handled by:
> >>>> drivers/platform/x86/i2c-multi-instantiate.c
> >>>>
> >>>> Which I tried to do at first, but as explained in the commit msg
> >>>> of the updated patch, that is not possible in this special case
> >>>> (because it would cause userspace breakage due to the modalias changing).    
> >>>
> >>> This is marked as patch 1?!    
> >>
> >> Yes my bad, sorry.
> >>  
> >>> Usually --cover-letter produces a correct template...    
> >>
> >> I use --compose, time to switch to --cover-letter instead I guess.
> >>
> >> Regards,
> >>
> >> Hans
> >>  
> > 
> > Other than the minor bits Andy mentioned I'm fine with this.
> > 
> > It does sort of feel like we should be gathering documentation somewhere of
> > where this crazy stuff has been seen in the wild.  
> 
> I can provide an (incomplete) list of devices known to use the BOSC0200 ACPI HID,
> both in single and dual accelerometer configs.
> 
> In lieu of a better place, I guess I could best just drop this info in
> a big comment block near the ACPI mount-matrix parsing stuff ?
> 
> If you agree that a comment is a reasonable place to place this info I
> can add this for v2 of the patch-set.
It is as good a place as any so go for it.

thanks,

Jonathan

p.s. Mailing list is messing around today I think as neither I nor lore.kernel.org
are getting mails from it.  It'll probably be resolved shorty but in meantime
discussion may get a bit separated and I can't use b4 for new patches which
will make it trickier to apply things until that is fixed.

> 
> Regards,
> 
> Hans
>