RE: [PATCH v2 1/3] ACPI: Resolve objects on host-directed table loads

[Nikolaus Voss]

On Wed, 19 Jun 2019, Moore, Robert wrote:
> > -----Original Message-----
> > From: Nikolaus Voss [mailto:nv@xxxxxxx]
> > Sent: Wednesday, June 19, 2019 2:31 AM
> > To: Moore, Robert <robert.moore@xxxxxxxxx>
> > Cc: Rafael J. Wysocki <rafael@xxxxxxxxxx>; Rafael J. Wysocki
> > <rjw@xxxxxxxxxxxxx>; Len Brown <lenb@xxxxxxxxxx>; Schmauss, Erik
> > <erik.schmauss@xxxxxxxxx>; Jacek Anaszewski
> > <jacek.anaszewski@xxxxxxxxx>; Pavel Machek <pavel@xxxxxx>; Dan Murphy
> > <dmurphy@xxxxxx>; Thierry Reding <thierry.reding@xxxxxxxxx>; ACPI Devel
> > Maling List <linux-acpi@xxxxxxxxxxxxxxx>; open list:ACPI COMPONENT
> > ARCHITECTURE (ACPICA) <devel@xxxxxxxxxx>; linux-leds@xxxxxxxxxxxxxxx;
> > Linux PWM List <linux-pwm@xxxxxxxxxxxxxxx>; Linux Kernel Mailing List
> > <linux-kernel@xxxxxxxxxxxxxxx>; nikolaus.voss@xxxxxxxxxxxxxxxxxxxxx
> > Subject: RE: [PATCH v2 1/3] ACPI: Resolve objects on host-directed table
> > loads
> >
> > Hi Bob,
> >
> > On Tue, 18 Jun 2019, Moore, Robert wrote:
> > > > -----Original Message-----
> > > > From: Moore, Robert
> > > > Sent: Tuesday, June 18, 2019 1:25 PM
> > > > To: 'Nikolaus Voss' <nv@xxxxxxx>
> > > > Cc: 'Rafael J. Wysocki' <rafael@xxxxxxxxxx>; 'Rafael J. Wysocki'
> > > > <rjw@xxxxxxxxxxxxx>; 'Len Brown' <lenb@xxxxxxxxxx>; Schmauss, Erik
> > > > <erik.schmauss@xxxxxxxxx>; 'Jacek Anaszewski'
> > > > <jacek.anaszewski@xxxxxxxxx>; 'Pavel Machek' <pavel@xxxxxx>; 'Dan
> > > > Murphy' <dmurphy@xxxxxx>; 'Thierry Reding'
> > > > <thierry.reding@xxxxxxxxx>; 'ACPI Devel Maling List'
> > > > <linux-acpi@xxxxxxxxxxxxxxx>; 'open list:ACPI COMPONENT ARCHITECTURE
> > > > (ACPICA)' <devel@xxxxxxxxxx>; 'linux- leds@xxxxxxxxxxxxxxx' <linux-
> > leds@xxxxxxxxxxxxxxx>; 'Linux PWM List'
> > > > <linux-pwm@xxxxxxxxxxxxxxx>; 'Linux Kernel Mailing List' <linux-
> > > > kernel@xxxxxxxxxxxxxxx>
> > > > Subject: RE: [PATCH v2 1/3] ACPI: Resolve objects on host-directed
> > > > table loads
> > > >
> > > > If it is in fact the AcpiLoadTable interface that is incorrect, that
> > > > of course is different. I'll check that out next.
> > > [Moore, Robert]
> > >
> > > Yes, this is the issue, not specifically the Load() operator, but the
> > > AcpiLoadTable interface only.
> >
> > thanks for checking this out. So what is the conclusion? Is my fix of
> > AcpiLoadTable() sufficient or do we need a different solution?
> >
> > Niko
>
>
> Your change is in the correct area. We want to do something like this, however:
>
> diff --git a/source/components/executer/exconfig.c b/source/components/executer/exconfig.c
> index 84a058ada..eba1a6d28 100644
> --- a/source/components/executer/exconfig.c
> +++ b/source/components/executer/exconfig.c
> @@ -342,10 +342,9 @@ AcpiExLoadTableOp (
>         return_ACPI_STATUS (Status);
>     }
>
> -    /* Complete the initialization/resolution of package objects */
> +    /* Complete the initialization/resolution of new objects */
>
> -    Status = AcpiNsWalkNamespace (ACPI_TYPE_PACKAGE, ACPI_ROOT_OBJECT,
> -        ACPI_UINT32_MAX, 0, AcpiNsInitOnePackage, NULL, NULL, NULL);
> +    AcpiNsInitializeObjects ();
>
>     /* Parameter Data (optional) */
>
> @@ -620,10 +619,11 @@ AcpiExLoadOp (
>         return_ACPI_STATUS (Status);
>     }
>
> -    /* Complete the initialization/resolution of package objects */
> +    /* Complete the initialization/resolution of new objects */
>
> -    Status = AcpiNsWalkNamespace (ACPI_TYPE_PACKAGE, ACPI_ROOT_OBJECT,
> -        ACPI_UINT32_MAX, 0, AcpiNsInitOnePackage, NULL, NULL, NULL);
> +    AcpiExExitInterpreter ();
> +    AcpiNsInitializeObjects ();
> +    AcpiExEnterInterpreter ();
>
>     /* Store the DdbHandle into the Target operand */
>
> diff --git a/source/components/tables/tbxfload.c b/source/components/tables/tbxfload.c
> index 217d54bf0..1e17db6c8 100644
> --- a/source/components/tables/tbxfload.c
> +++ b/source/components/tables/tbxfload.c
> @@ -479,6 +479,13 @@ AcpiLoadTable (
>     ACPI_INFO (("Host-directed Dynamic ACPI Table Load:"));
>     Status = AcpiTbInstallAndLoadTable (ACPI_PTR_TO_PHYSADDR (Table),
>         ACPI_TABLE_ORIGIN_EXTERNAL_VIRTUAL, FALSE, &TableIndex);
> +    if (ACPI_SUCCESS (Status))
> +    {
> +        /* Complete the initialization/resolution of new objects */
> +
> +        AcpiNsInitializeObjects ();
> +    }
> +
>     return_ACPI_STATUS (Status);
> }

Ok, I see your are taking this up (I was a bit unsure after your previous 
post). Thanks,
Niko