Re: [PATCH 70/73] ACPICA: Fix to disable unknown spurious GPEs

[Alexey Starikovskiy <astarikovskiy@suse.de>]

Actually, I too don't know why _disabling_ some GPE could trigger EC one...
May be Len could try to debug this issue, as he is the only one with hardware...
One thought is that EC GPE is not present in GPE scope, thus it will be not configured by the 
time 6217' code is executed. So, in theory, 6217 could try to disable EC GPE.

Regards,
Alex.


Moore, Robert wrote:
> Yes, this is my question also.
> 
> I hope Alexey can help with this. I'm very nervous about any code that
> works or does not work for reasons unknown. When this happens in the GPE
> code, I get more than nervous.
> 
> Bob
> 
> 
> -----Original Message-----
> From: Zhang, Rui 
> Sent: Thursday, April 24, 2008 6:53 PM
> To: Len Brown
> Cc: Moore, Robert; Alexey Starikovskiy; linux-acpi@vger.kernel.org
> Subject: Re: [PATCH 70/73] ACPICA: Fix to disable unknown spurious GPEs
> 
> 
> On Fri, 2008-04-25 at 09:40 +0800, Len Brown wrote:
>> oh goody, i thought this one rang a bell...
>>
>> I'll apply 6217 on top of 9916 (which was already in the test tree,
>> but not on the acpica branch)
>>
>> the error goes away with both applied.
> 
> I still don't see why the error messages pop up if 6217 is applied.
> IMO, 6217 won't enable the EC GPE. :(
> 
> thanks,
> rui
> 
>> On Wednesday 23 April 2008, Moore, Robert wrote:
>>> Alexey,
>>>
>>> I remember there was some discussion on this, but I'm still not sure
>> why
>>> the patch for 6217 (Ignore GPE if valid and not enabled) makes the
>> other
>>> problem (too early enabling of GPEs) apparent. Please explain.
>>>
>>> Thanks,
>>> Bob
>>>
>>>
>>>> -----Original Message-----
>>>> From: Alexey Starikovskiy [mailto:astarikovskiy@suse.de]
>>>> Sent: Tuesday, April 22, 2008 10:29 PM
>>>> To: Len Brown
>>>> Cc: linux-acpi@vger.kernel.org; Moore, Robert; Zhang, Rui
>>>> Subject: Re: [PATCH 70/73] ACPICA: Fix to disable unknown spurious
>> GPEs
>>>> Len Brown wrote:
>>>>> On Saturday 12 April 2008, Len Brown wrote:
>>>>>> From: Bob Moore <robert.moore@intel.com>
>>>>>>
>>>>>> Implemented another change to eliminate/suppress spurious or
>>>>>> stray GPEs. The AcpiEvDisableGpe function will now disable GPEs
>>>>>> that are neither enabled nor disabled -- meaning that the GPE is
>>>>>> unknown to the system. This will prevent future interrupts from
>>>>>> that GPE. (Zhang Rui) BZ 6217
>>>>>>
>>>>>> Signed-off-by: Bob Moore <robert.moore@intel.com>
>>>>>> Signed-off-by: Alexey Starikovskiy <astarikovskiy@suse.de>
>>>>>> Signed-off-by: Len Brown <len.brown@intel.com>
>>>>>> ---
>>>>>>  drivers/acpi/events/evgpe.c |    9 ++++++++-
>>>>>>  1 files changed, 8 insertions(+), 1 deletions(-)
>>>>>>
>>>>>> diff --git a/drivers/acpi/events/evgpe.c
>>> b/drivers/acpi/events/evgpe.c
>>>>>> index 897baed..6940f5d 100644
>>>>>> --- a/drivers/acpi/events/evgpe.c
>>>>>> +++ b/drivers/acpi/events/evgpe.c
>>>>>> @@ -250,7 +250,14 @@ acpi_status acpi_ev_disable_gpe(struct
>>>> acpi_gpe_event_info *gpe_event_info)
>>>>>>   ACPI_FUNCTION_TRACE(ev_disable_gpe);
>>>>>>
>>>>>> - if (!(gpe_event_info->flags & ACPI_GPE_ENABLE_MASK)) {
>>>>>> + /*
>>>>>> +  * Ignore this if the GPE is valid and not enabled.
>>>>>> +  *
>>>>>> +  * Flags is only zero if GPE is neither enabled or disabled --
>>> it may
>>>>>> +  * be a spurious or stray GPE -- disable it in the default
>> case
>>>> below.
>>>>>> +  */
>>>>>> + if (gpe_event_info->flags &&
>>>>>> +     (!(gpe_event_info->flags & ACPI_GPE_ENABLE_MASK))) {
>>>>>>           return_ACPI_STATUS(AE_OK);
>>>>>>   }
>>>>>>
>>>>>
>>>>> When applied to 2.6.25, the patch above causes ACPI Error
>> messages
>>> below
>>>> on my T61.
>>>>> Certainly this is a case of the EC getting an interrupt (GPE x02)
>>> before
>>>> it has initialized.
>>>>> Perhaps it breaks the call of acpi_install_gpe_handler() to
>>>> acpi_ev_disable_gpe()?
>>>> Congratulations, now you too hit this one, patch is already
>> there...
>>>> It is wroong to enable GPE before handler for it is installed...
>>>> http://bugzilla.kernel.org/show_bug.cgi?id=9916
>>>>
>>>>> -Len
>>>>>
>>>>>
>>>>>  PCI: Using MMCONFIG at f0000000 - f3ffffff
>>>>>  PCI: Using configuration type 1
>>>>>  evgpeblk-0956 [00] ev_create_gpe_block   : GPE 00 to 1F [_GPE] 4
>>> regs on
>>>> int 0x9
>>>>> +ACPI Error (evregion-0316): No handler for Region [ECOR]
>>>> (ffff81007d35dda0) [EmbeddedControl] [20070126]
>>>>> +ACPI Error (exfldio-0289): Region EmbeddedControl(3) has no
>> handler
>>>> [20070126]
>>>>> +ACPI Error (psparse-0537): Method parse/execution failed
>>>> [\_SB_.PCI0.LPC_.EC__.AC__._PSR] (Node ffff81007d364ab0),
>> AE_NOT_EXIST
>>>>> +ACPI Error (psparse-0537): Method parse/execution failed
>>>> [\_SB_.PCI0.LPC_.EC__.HKEY.CKC4] (Node ffff81007d364f48),
>> AE_NOT_EXIST
>>>>> +ACPI Error (psparse-0537): Method parse/execution failed
>>> [\_GPE._L02]
>>>> (Node ffff81007d376420), AE_NOT_EXIST
>>>>> +ACPI Exception (evgpe-0584): AE_NOT_EXIST, while evaluating GPE
>>> method
>>>> [_L02] [20070126]
>>>>>  evgpeblk-1052 [00] ev_initialize_gpe_bloc: Found 9 Wake, Enabled
>> 3
>>>> Runtime GPEs in this block
>>>>>  ACPI: EC: EC description table is found, configuring boot EC
>>>>> -ACPI: EC: non-query interrupt received, switching to interrupt
>> mode
>>>>> -Completing Region/Field/Buffer/Package
>> initialization:........................................................
>>> ....
>>>> ............................................................
>> .......................................................................
>>> ....
>>>
>> .......................................................................
>>> ....
>>>> ................
>>>>> -Initialized 29/33 Regions 138/138 Fields 64/64 Buffers 55/72
>>> Packages
>>>> (1888 nodes)
>>>>> +Completing Region/Field/Buffer/Package
>> initialization:........................................................
>>> ....
>>>> ............................................................
>> .......................................................................
>>> ....
>>>
>> .......................................................................
>>> ....
>>>> .................
>>>>> +Initialized 30/33 Regions 138/138 Fields 64/64 Buffers 55/72
>>> Packages
>>>> (1888 nodes)
>>>>>  Initializing Device/Processor/Thermal objects by executing _INI
>>>> methods:.<5>ACPI: BIOS _OSI(Linux) query honored via DMI
>>>>> +ACPI: EC: non-query interrupt received, switching to interrupt
>> mode
>>>>>  .......
>>>>>  Executed 8 _INI methods requiring 2 _STA executions (examined 88
>>>> objects)
>>>>>  ACPI: Interpreter enabled
>>>
>>
>>
>>
>