ACPI 5.0

[Alex Hung]

Hi,

I quickly read through the revision history and identify what's new in 
ACPI 5.0, as summarized below:

Any comments are most welcomed.

== Summary ==

Many new changes are related to ARM platform, and may not related to 
current x86 designs.
However, we can still check new acpi tables and control methods whether 
they are used in ARM or in PC as this is not particularly difficult. I 
am guessing it won't be long that we see ARM features on x86.

Ubuntu on ARM does not support on ACPI (?), but this may change with the 
introduction of Win8 on ARM. I suggest to lower the priority for 
ARM-specific changes that will consume much time, but may need to finish 
them in ubuntu-r cycle.

Does anyone know the ubuntu's plan for ACPI on ARM?

== HP-0001 ==

  * 5.6.6 Device Object Notifications - Graceful Shutdown Request (0x0c)
  * Need to check whether linux kernel supports it
   * probably no need to test it in fwts (or check it at particular 
methods)

== INTC-001 ==

  * Firmware Performance Data Table (FPDT) - used to describe the 
platform initialization performance records
   * used to show boot time
   * fwts needs to check it
    * can also decode it for further suggestions

== INTC-002 ==

  * Error Injection Table (EINJ) - for OSPM can inject hardware errors 
to the platform
   * it exists in ACPI 4.0a
   * fwts needs to check it (if not done so yet)

== INTC-003 ==

  * Memory Power StateTable (MPST) - defines the memory power node 
topology of the configuration
   * fwts needs to check it

== INTC-008 & INTC-009 ==

  * ACPI RAS FeatureTable (RASF)
   * fwts needs to check it

== MSFT-0001 ==

  * HW-reduced ACPI is for ARM
   * fwts needs to check it in long term

== MSFT-0002 ==

  * Generic interrupt Controller is used on ARM
   * fwts needs to check it in long term

== MSFT-0003 ==

  * Device Identification Objects
   * new objects: _CLS, _HRV and _SUB
    * fwts needs to check them

== MSFT-0004 ==

  * GPIO Abstraction - new device type: GPIO controller
   * new methods: _AIE and _EVT
    * probably for ARM
    * fwts needs to check them
    * _EVT's input is odd - feedback to acpi sig

== MSFT-0006 ==

  * Simple Peripheral Bus (SPB)
   * mostly used on ARM?
   * new object: _DEP
    * fwts needs to check it

== MSFT-0007 ==

  * a new method: _CPC
   * fwts needs to check it

== MSFT-0008 ==

  * A section 14 for Platform Communications Channel (PCC)
   * a new table: Platform Communications Channel Table (PCCT)
   * fwts needs to check it
   * not sure whether it is for ARM or generic...

== MSFT-0009 ==

  * Time and Alarm Device <- originally it was Wake Alarm Device
   * new methods: _GCP, _GRT, _SRT, _GWS, _CWS, _STP
    * fwts needs to check them

== MSFT-0014 ==

  * Firmware Performance Data Table (FPDT) - see INTC-001

== MSFT-0016 ==

  * see MSFT-0004

== MSFT-0018 ==

  * a new method: _DLM
   * fwts needs to check it
   * does linux kernel support it?

== MSFT-0019 ==

  * Generic Timer Description Table (GTDT)
   * for ARM
   * fwts needs to check it

== MSFT-0020 ==

  * new methods: _PSR, _PRE
   * fwts needs to check them


On 07/06/2012 04:03 PM, Colin Ian King wrote:
> Hi fwts developers,
>
> We are en-route to making fwts handle ACPI 5.0, but there is still a
> *lot* more work to be done.
>
> Here are some tasks that need completing:
>
> 1) Check out the ACPI 5.0 spec [1] revision history (pages iii, iv) to
> see what's new.
>
> 2) Identify new ACPI tables and add support for them (e.g. add to fwts
> acpidump and where necessary add tests to verify these tables).  [ We
> may want to also identify tables missing from fwts for ACPI 4.0 too ]
>
> 3) Identify new ACPI 5.0 controls and add tests for these to the
> "method" test.  I've already enumerated all of these new controls (and
> found a few that I omitted from ACPI 4.0) but we need still need to
> write a whole load of tests for these.
>
> 4) Identify changes to existing tables (added fields + features) and
> expand existing tests where necessary.
>
> As a sanity check, we should also look at the recent kernel changes to
> the ACPI driver see what's new, learn from this and see if there are any
> new sanity checks we should be putting into fwts. This breaks down to:
>
> a) Do we need to completely write new tests?
> b) Do we need to extend the current tests? e.g. battery tests, klog,
> fadt, etc.
>
> ..so lots more to do.
>
> I propose we get somebody to work through the list of changes and write
> these up.  We should then prioritize the changes to fwts to cater for
> these changes. Then we can get volunteers to opt-in to implement the
> missing features.  If one is starting to work on a feature it makes
> sense to inform others in this mailing list.
>
> Anyhow, I'm open for discussion on how to move this forward and for any
> suggestions to help make this plan work.
>
> Discuss! :-)
>
> Colin
>
> References:
>
> [1] http://www.acpi.info/DOWNLOADS/ACPIspec50.pdf
>