Re: [PATCH v6 00/12] Add ACPI _DSD and unified device properties support

["Rafael J. Wysocki" <rjw@rjwysocki.net>]

On Tuesday, October 21, 2014 11:08:59 PM Rafael J. Wysocki wrote:
> Hi Everyone,
> 
> This is version 6 of the unified device properties interface patchset.
> 
> The original cover letter from Mika is here:
> 
> http://marc.info/?l=devicetree&m=141087052200600&w=4
> 
> and my cover letters for previous iterations are at:
> 
> http://marc.info/?l=linux-acpi&m=141212903816560&w=4
> http://marc.info/?l=linux-kernel&m=141354745011569&w=4
> 
> There are a few changes with respect to v5 and the affected patches are
> [02-03/12] and [09-12/12].  The remaining ones have not been modified.
> 
> Most importantly, requesting the first element of a list (package) property
> from _DSD is now equivalent to accessing a single-value property of the
> same type, so device_property_read_u8(dev, pname, val) will now be equivalent
> to device_property_read_u8_array(dev, pname, val, 1), for example.
> Consequently, this _DSD definition:
> 
> Name (_DSD, Package () {
>     ToUUID("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"),
>     Package () {
>         Package () {"blah", "A string"},
>     }
> })
> 
> can be used instead of
> 
> Name (_DSD, Package () {
>     ToUUID("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"),
>     Package () {
>         Package () {"blah", Package () {"A string"}},
>     }
> })
> 
> and the code will be able to retrieve the property value from the both of
> them just fine.
> 
> This means, among other things, that accessors for single-value properties
> can be implemented in terms of the analogous "array" property accessors
> which allows the code size to be reduced somewhat.
> 
> Patches [02/12] and [09/12] have been modified to achieve that and patch
> [03/12] have been modified accordingly for the "compatible" property in
> _DSD to behave in an analogous way.  Additionally, the bodies of the
> numerical property accessors in patches [02/12] and [09/12] are now
> generated using macros (string property accessors have slightly different
> rules and are simply open coded for that reason).
> 
> Patch [10/12] has been modified to drop function arguments that happened to
> have the same values for both of the current users of those functions and
> patches [11-12/12] have been modified to take that change into account.  If
> the code in question needs to be made more complex in the future, there
> should not be any problems with that.
> 
> Due to the nature of the changes I have retained all ACKs except for the
> Grant's Reviewed-by on patch [03/12] (if that had been Acked-by, I would have
> retained it too, but that didn't feel appropriate for the "reviewed by" thing
> to me).  If any of you think that the ACKs are not applicable any more, please
> let me know and I'll drop them.
> 
> Finally, many thanks to Mika for testing the series on MinnowBoard 1 and
> MinnowBoard Max.  In case anybody else would like to test it, it is available
> from the device-properties branch of the linux-pm.git tree:
> 
> git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git device-properties
> 
> Thanks!

Crickets ...

OK, so I'm taking the lack of comments as the lack of objections and I'm already
getting merge conflicts for this series.  Moreover, we already have done some
work on top of it.

So, if there still are no comments by Sunday evening, I'll add this series to
my linux-next branch with 3.19-rc1 as the target.

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.