Re: [Qemu-devel] [PATCH v14 7/9] ARM: ACPI: Add GPIO notification type for hardware RAS error

[Igor Mammedov <imammedo@redhat.com>]

On Wed, 3 Jan 2018 11:48:30 +0800
gengdongjiu <gengdongjiu@huawei.com> wrote:

> On 2017/12/28 22:53, Igor Mammedov wrote:
> > On Thu, 28 Dec 2017 13:54:16 +0800
> > Dongjiu Geng <gengdongjiu@huawei.com> wrote:
[...]
> >> +static void acpi_dsdt_add_error_device(Aml *scope)
> >> +{
> >> +    Aml *dev = aml_device(ACPI_HARDWARE_ERROR_DEVICE);
> >> +    Aml *method;
> >> +
> >> +    aml_append(dev, aml_name_decl("_HID", aml_eisaid("PNP0C33")));
> >> +    aml_append(dev, aml_name_decl("_UID", aml_int(0)));
> >> +
> >> +    method = aml_method("_STA", 0, AML_NOTSERIALIZED);
> >> +    aml_append(method, aml_return(aml_int(0x0f)));  
> > no need for dummy _STA method, device is assumed to be present if there is no _STA   
> Igor,
>   do you mean remove above two line code as shown in [1]?
> I dump the DSDT table in my host Ubuntu PC for the error device (PNP0C33), it has the _STA, as shown in [2].
> do we not want to add the _STA for guest?
> 
> [1]
> +    method = aml_method("_STA", 0, AML_NOTSERIALIZED);
> +    aml_append(method, aml_return(aml_int(0x0f)));
compared to host, yours method does nothing,
read ACPI6.2 "6.3.7 _STA (Status)" one more time

> [2]:
>         Device (WERR)
>         {
>             Name (_HID, EisaId ("PNP0C33"))  // _HID: Hardware ID
>             Method (_STA, 0, NotSerialized)  // _STA: Status
>             {
>                 If (LGreaterEqual (OSYS, 0x07D9))
>                 {
>                     Return (0x0F)
>                 }
>                 Else
>                 {
>                     Return (Zero)
>                 }
>             }
>         }
> >   
> >> +    aml_append(dev, method);
> >> +    aml_append(scope, dev);
> >> +}
> >> +
[...]