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

[Nikolaus Voss <nv@vosn.de>]

On Mon, 17 Jun 2019, Moore, Robert wrote:
>> -----Original Message-----
>> From: Nikolaus Voss [mailto:nv@vosn.de]
>> Sent: Sunday, June 16, 2019 11:24 PM
>> To: Moore, Robert <robert.moore@intel.com>
>> Cc: Rafael J. Wysocki <rafael@kernel.org>; Rafael J. Wysocki
>> <rjw@rjwysocki.net>; Len Brown <lenb@kernel.org>; Schmauss, Erik
>> <erik.schmauss@intel.com>; Jacek Anaszewski
>> <jacek.anaszewski@gmail.com>; Pavel Machek <pavel@ucw.cz>; Dan Murphy
>> <dmurphy@ti.com>; Thierry Reding <thierry.reding@gmail.com>; ACPI Devel
>> Maling List <linux-acpi@vger.kernel.org>; open list:ACPI COMPONENT
>> ARCHITECTURE (ACPICA) <devel@acpica.org>; linux-leds@vger.kernel.org;
>> Linux PWM List <linux-pwm@vger.kernel.org>; Linux Kernel Mailing List
>> <linux-kernel@vger.kernel.org>; nikolaus.voss@loewensteinmedical.de
>> Subject: RE: [PATCH v2 1/3] ACPI: Resolve objects on host-directed table
>> loads
>>
>> Bob,
>>
>> On Fri, 14 Jun 2019, Moore, Robert wrote:
>>>
>>>
>>> -----Original Message-----
>>> From: Nikolaus Voss [mailto:nv@vosn.de]
>>> Sent: Friday, June 14, 2019 2:26 AM
>>> To: Rafael J. Wysocki <rafael@kernel.org>
>>> Cc: Rafael J. Wysocki <rjw@rjwysocki.net>; Len Brown
>>> <lenb@kernel.org>; Moore, Robert <robert.moore@intel.com>; Schmauss,
>>> Erik <erik.schmauss@intel.com>; Jacek Anaszewski
>>> <jacek.anaszewski@gmail.com>; Pavel Machek <pavel@ucw.cz>; Dan Murphy
>>> <dmurphy@ti.com>; Thierry Reding <thierry.reding@gmail.com>; ACPI
>>> Devel Maling List <linux-acpi@vger.kernel.org>; open list:ACPI
>>> COMPONENT ARCHITECTURE (ACPICA) <devel@acpica.org>;
>>> linux-leds@vger.kernel.org; Linux PWM List
>>> <linux-pwm@vger.kernel.org>; Linux Kernel Mailing List
>>> <linux-kernel@vger.kernel.org>
>>> Subject: Re: [PATCH v2 1/3] ACPI: Resolve objects on host-directed
>>> table loads
>>>
>>> Hi Rafael,
>>>
>>> On Fri, 14 Jun 2019, Rafael J. Wysocki wrote:
>>>> On Wed, Jun 12, 2019 at 10:36 AM Nikolaus Voss
>>>> <nikolaus.voss@loewensteinmedical.de> wrote:
>>>>>
>>>>> If an ACPI SSDT overlay is loaded after built-in tables have been
>>>>> loaded e.g. via configfs or efivar_ssdt_load() it is necessary to
>>>>> rewalk the namespace to resolve references. Without this, relative
>>>>> and absolute paths like ^PCI0.SBUS or \_SB.PCI0.SBUS are not
>>>>> resolved correctly.
>>>>>
>>>>> Make configfs load use the same method as efivar_ssdt_load().
>>>>>
>>>>> Signed-off-by: Nikolaus Voss <nikolaus.voss@loewensteinmedical.de>
>>>>
>>>> This is fine by me, so
>>>>
>>>> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>>>>
>>>> Or if you want me to take this patch (without the other two in the
>>>> series), please let me know.
>>>
>>> thanks. I think it would be the best if you take up this patch as it
>>> is an independent topic. In retrospect it wasn't a good idea to put it
>>> into this series.
>>>
>>> Kind regards,
>>> Niko
>>>
>>> I would have to ask, why is additional code needed for package
>>> initialization/resolution? It already happens elsewhere in acpica. Bob
>>
>> for built-in tables loaded via acpi_ex_load_table_op() everything is
>> fine, because after acpi_tb_load_table() acpi_ns_walk_namespace() is
>> called to resolve references.
>>
>> My fix only affects tables loaded dynamically via either acpi_configfs
>> driver (for debugging purposes) or efivar_ssdt_load() (to specify a
>> table on the kernel's command line). They use acpi_load_table() to load
>> the table from a caller-owned buffer. To resolve the references, it is
>> again necessary to rewalk the namespace, which was simply missing in
>> acpi_load_table().
>>
> [Moore, Robert]
>
> Perhaps you should call AcpiInitializeObjects after the call to 
> AcpiLoadTable, but I will check.

My usage of acpi_load_table() is to load a SSDT which is the intended use 
of this method according to its description. And my expectation is that 
the package objects of the loaded table are initialized when this function 
successfully returns so it aligns with the behavior of 
acpi_ex_load_table_op() for built-in SSDTs. Otherwise there would be no 
point in having this function at all, because 
acpi_tb_install_and_load_table() could be called directly without any 
difference.

Niko