Re: [V8 PATCH 1/3] ACPICA: Add ACPI _CLS processing

[Suravee Suthikulpanit]

On 04/27/2015 03:38 PM, Rafael J. Wysocki wrote:
> On Sunday, April 26, 2015 10:45:29 PM Suthikulpanit, Suravee wrote:
> > On 4/24/15, 21:28, "Rafael J. Wysocki" <rjw@xxxxxxxxxxxxx> wrote:
> >
> > > On Friday, April 24, 2015 04:08:31 PM Suravee Suthikulpanit wrote:
> > > > On 4/16/15 20:45, Zheng, Lv wrote:
> > > > > Before back porting this to ACPICA, let me ask one simple question.
> > > > > According to the spec, the _CLS is optional and PCI specific.
> > > > > So why should we implement it in ACPICA core not OSPM specific
> > > > modules?
> > > > > If this need to be implemented in ACPICA, then what about the
> > > > following device identification objects?
> > > > > _DDN, _HRV, _MLS, _PLD, _STR, _SUN
> > > > >
> > > > > Thanks and best regards
> > > > > -Lv
> > > >
> > > > Hi,
> > > >
> > > > Sorry for late reply. As for the justification for introducing the _CLS
> > > > support in the ACPICA, this is mainly because ACPI does not currently
> > > > define _CID for certain device classes, which used to mostly be PCI
> > > > devices. Instead, ACPI spec mentioned that _CLS can be used for loading
> > > > generic drivers on hardware that is compatible with PCI-defined device
> > > > classes, but that is not implemented on the PCI bus (and is therefore
> > > > enumerated by ACPI.)
> > >
> > > I think it would be good to point to the particular part of the spec
> > > making that provision.  In what section is that mentioned, exactly?
> >
> > Here is the copied from section 6.1.3 _CLS (Class Code) from ACPI 5.1 spec:
> > "This object is used to supply OSPM with the PCI-defined class, subclass
> > and programming interface for a device. This object is optional but may be
> > useful for generic drivers written for PCI devices that move off of PCI
> > and are enumerated by ACPI."
>
> OK, so the "move off of PCI" part should be understood as "something that
> used to be on the PCI bus, but now may be included into an SoC directly
> in which case it won't be a PCI device any more", right?

That's right. For example, the SATA controller is a good example for 
this case. On most x86 platforms, they are often enumerated as PCI 
devices. However, in the ARM64 SOC (e.g. on AMD Seattle), it could be 
enumerated as non-PCI device.

Thanks,

Suravee

> > Otherwise, if the community think itÂs better to not putting the _CLS the
> > _CLS parsing code in ACPICA since, I can try looking into pulling the code
> > out of ACPICA. I also noticed a little issue in the patch series where the
> > ACPI_VALID_CLS is used in the patch 1 but defined in patch 2. I can send
> > out v9 with the fix once we agree on the _CLS parsing.
>
> To me that pretty much depends on the answer to the above question.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/