Re: [PATCH 0/4] Fixups for s2idle on ASUS Rembrandt laptop

[Mario Limonciello <mario.limonciello@amd.com>]

On 9/18/22 17:17, Luke Jones wrote:
> On Sun, 2022-09-18 at 14:46 -0500, Mario Limonciello wrote:
>> On 9/18/22 03:37, Luke Jones wrote:
>>> Hi,
>>>
>>> On Mon, 2022-09-12 at 17:06 +0200, Philipp Zabel wrote:
>>>> Am Mon, Sep 12, 2022 at 02:58:51PM +0000 schrieb Limonciello,
>>>> Mario:
>>>>> [AMD Official Use Only - General]
>>>>>
>>>>>
>>>>>
>>>>>> -----Original Message-----
>>>>>> From: Philipp Zabel <philipp.zabel@gmail.com>
>>>>>> Sent: Monday, September 12, 2022 09:57
>>>>>> To: Limonciello, Mario <Mario.Limonciello@amd.com>
>>>>>> Cc: rafael@kernel.org; catalin@antebit.com; S-k, Shyam-sundar
>>>>>> <Shyam-
>>>>>> sundar.S-k@amd.com>; Len Brown <lenb@kernel.org>; linux-
>>>>>> acpi@vger.kernel.org; linux-kernel@vger.kernel.org
>>>>>> Subject: Re: [PATCH 0/4] Fixups for s2idle on ASUS Rembrandt
>>>>>> laptop
>>>>>>
>>>>>> Hi Mario,
>>>>>>
>>>>>> Am Fri, Sep 09, 2022 at 01:05:05PM -0500 schrieb Mario
>>>>>> Limonciello:
>>>>>>> It was reported that an ASUS Rembrandt laptop has problems
>>>>>>> with
>>>>>> seemingly
>>>>>>> unrelated ACPI events after resuming from s2idle. Debugging
>>>>>>> the
>>>>>>> issue
>>>>>>> proved it's because ASUS has ASL that is only called when
>>>>>>> using
>>>>>>> the
>>>>>>> Microsoft GUID, not the AMD GUID.
>>>>>>>
>>>>>>> This is a bug from ASUS firmware but this series reworks
>>>>>>> the
>>>>>>> s2idle
>>>>>>> handling for AMD to allow accounting for this in a quirk.
>>>>>>>
>>>>>>> Additionally as this is a problem that may pop up again on
>>>>>>> other models
>>>>>>> add a module parameter that can be used to try the
>>>>>>> Microsoft
>>>>>>> GUID on a
>>>>>>> given system.
>>>>>>
>>>>>> thank you, these also helped on an ASUS ROG Zephyrus G14
>>>>>> (2022)
>>>>>> with
>>>>>> BIOS version GA402RJ.313. Patches 1-3
>>>>>>
>>>>>> Tested-by: Philipp Zabel <philipp.zabel@gmail.com> # GA402RJ
>>>>>
>>>>> Did you use acpi.prefer_microsoft_guid=1 for your system then
>>>>> too?
>>>>>
>>>>> If so, I should re-spin this series to add your system's quirk
>>>>> to
>>>>> patch 4.
>>>>
>>>> Yes. I also tested with the following diff applied instead of the
>>>> module
>>>> parameter:
>>>>
>>>> ----------8<----------
>>>> diff --git a/drivers/acpi/x86/s2idle.c
>>>> b/drivers/acpi/x86/s2idle.c
>>>> index 6a2c94fdbeae..a247560e31de 100644
>>>> --- a/drivers/acpi/x86/s2idle.c
>>>> +++ b/drivers/acpi/x86/s2idle.c
>>>> @@ -420,6 +420,14 @@ static const struct dmi_system_id
>>>> s2idle_dmi_table[] __initconst = {
>>>>                           DMI_MATCH(DMI_PRODUCT_NAME, "ASUS TUF
>>>> Gaming
>>>> A17"),
>>>>                   },
>>>>           },
>>>> +       {
>>>> +               /* ASUS ROG Zephyrus G14 (2022) */
>>>> +               .callback = lps0_prefer_microsoft,
>>>> +               .matches = {
>>>> +                       DMI_MATCH(DMI_SYS_VENDOR, "ASUSTeK
>>>> COMPUTER
>>>> INC."),
>>>> +                       DMI_MATCH(DMI_PRODUCT_NAME, "ROG Zephyrus
>>>> G14
>>>> GA402"),
>>>
>>> Just a note, this needs to be `ROG Zephyrus G14 GA402R` or this
>>> will
>>> catch the GA402Q series as well which doesn't require this quirk.
>>> In general the model numbering goes
>>> <Range><Model><Generation><Graphics>, So for my old G14
>>> <GA><402<Q><M>.
>>> Or for example a ROG Strix machine <G><513><Q><Y>
>>>
>>
>> What _HID is used?  From your description GA402"Q" is Cezanne
>> generation
>> right? Can you please share the acpidump for me to confirm what is
>> happening and if it's expected?
> 
> Sure, dumps are here -
> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitlab.com%2Fasus-linux%2Freverse-engineering%2F-%2Ftree%2Fmaster%2Fga401qm%2F408-fw&amp;data=05%7C01%7Cmario.limonciello%40amd.com%7Cf7ff7320f72f483145c108da99c39de1%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637991362709739412%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=S8%2Bqp%2BtwFh6j1ZikKLpwRodmrhhSnthVDwuJZ8Ncs0A%3D&amp;reserved=0
> 
> That repo is a bit haphazard as info tends to get collected
> sporadically when issues arise, but it may be of use for you for other
> things. I'll try to clean it up some.
> 

Thanks for sharing it.

So this system uses "AMDI0005", so it matches "amd_picasso" behavior in 
the patch series.  That is "data->prefer_amd_guid" is false, so the 
quirk and parameter shouldn't actually do anything in this system.

>>
>>> I don't know of any others that may need this quirk.
>>>
>>>> +               },
>>>> +       },
>>>>           {}
>>>>    };
>>>>    
>>>> ---------->8----------
>>>>
>>>> The full DMI Product Name is "ROG Zephyrus G14 GA402RJ_GA402RJ",
>>>> but
>>>> there is also a near-identical higher spec model GA402RK.
>>>>
>>>> regards
>>>> Philipp
>>>
>>
>