Re: [PATCH 1/3] intel_pstate: skip the driver if Sun server has ACPI _PPC method

From: Linda Knippers
Date: Mon Nov 24 2014 - 10:54:29 EST


On 11/23/2014 8:41 PM, ethan zhao wrote:
> Linda,
>
> On 2014/11/21 12:44, Linda Knippers wrote:
>>
>> On 11/20/2014 07:37 PM, ethan zhao wrote:
>>> Dirk,
>>>
>>> On 2014/11/21 0:50, Dirk Brandewie wrote:
>>>> On 11/19/2014 12:22 PM, Linda Knippers wrote:
>>>>> On 11/18/2014 3:37 AM, Ethan Zhao wrote:
>>>>>> Oracle Sun X86 servers have dynamic power capping capability that
>>>>>> works via
>>>>>> ACPI _PPC method etc, so skip loading this driver if Sun server has
>>>>>> ACPI _PPC
>>>>>> enabled.
>>>>>>
>>>>>> Signed-off-by: Ethan Zhao <ethan.zhao@xxxxxxxxxx>
>>>>>> ---
>>>>>> drivers/cpufreq/intel_pstate.c | 20 ++++++++++++++++++++
>>>>>> 1 file changed, 20 insertions(+)
>>>>>>
>>>>>> diff --git a/drivers/cpufreq/intel_pstate.c
>>>>>> b/drivers/cpufreq/intel_pstate.c
>>>>>> index 27bb6d3..5498eb0 100644
>>>>>> --- a/drivers/cpufreq/intel_pstate.c
>>>>>> +++ b/drivers/cpufreq/intel_pstate.c
>>>>>> @@ -943,6 +943,21 @@ static bool intel_pstate_no_acpi_pss(void)
>>>>>> return true;
>>>>>> }
>>>>>>
>>>>>> +static bool intel_pstate_has_acpi_ppc(void)
>>>>>> +{
>>>>>> + int i;
>>>>>> +
>>>>>> + for_each_possible_cpu(i) {
>>>>>> + struct acpi_processor *pr = per_cpu(processors, i);
>>>>>> +
>>>>>> + if (!pr)
>>>>>> + continue;
>>>>>> + if (acpi_has_method(pr->handle, "_PPC"))
>>>>>> + return true;
>>>>>> + }
>>>>>> + return false;
>>>>>> +}
>>>>>> +
>>>>>> struct hw_vendor_info {
>>>>>> u16 valid;
>>>>>> char oem_id[ACPI_OEM_ID_SIZE];
>>>>>> @@ -952,6 +967,7 @@ struct hw_vendor_info {
>>>>>> /* Hardware vendor-specific info that has its own power management
>>>>>> modes */
>>>>>> static struct hw_vendor_info vendor_info[] = {
>>>>>> {1, "HP ", "ProLiant"},
>>>>>> + {1, "ORACLE", ""},
>>>>>> {0, "", ""},
>>>>>> };
>>>>>>
>>>>>> @@ -969,12 +985,16 @@ static bool
>>>>>> intel_pstate_platform_pwr_mgmt_exists(void)
>>>>>> !strncmp(hdr.oem_table_id, v_info->oem_table_id,
>>>>>> ACPI_OEM_TABLE_ID_SIZE) &&
>>>>>> intel_pstate_no_acpi_pss())
>>>>>> return true;
>>>>>> + if (!strncmp(hdr.oem_id, v_info->oem_id, ACPI_OEM_ID_SIZE) &&
>>>>>> + intel_pstate_has_acpi_ppc())
>>>>> We need try this on a few platforms to make sure this patch doesn't
>>>>> break the
>>>>> HP platforms that may or may not need this driver, depending on the
>>>>> BIOS settings.
>>>>>
>>>> It looks like HP systems would get swept up in this check too if they
>>>> have _PPC
>> Right. This patch breaks HP ProLiant platforms when they are
>> configured to have the OS do power management. In that case,
>> the firmware exposes _PPC information.
> Okay, got it, The HP ProLiant has an option in BIOS could be enabled to "OS
> PM", so
> will export _PSS, _PPC, and this patch break this case.
>
>>
>>> No , this patch checks the oem_id against 'ORACLE" first, will not
>>> affect other vendors even they have _PPC.
>> I don't think that's how your code works. This patch will match any
>> vendor that is in the table, not just "ORACLE".
> Will change patch to match the oem-id out of the loop, such as following , how
> about it ?
>
> static bool intel_pstate_platform_pwr_mgmt_exists(void)
> {
> struct acpi_table_header hdr;
> struct hw_vendor_info *v_info;
>
> if (acpi_disabled
> || ACPI_FAILURE(acpi_get_table_header(ACPI_SIG_FADT, 0, &hdr)))
> return false;
>
> for (v_info = vendor_info; v_info->valid; v_info++) {
> if (!strncmp(hdr.oem_id, v_info->oem_id, ACPI_OEM_ID_SIZE)
> && !strncmp(hdr.oem_table_id, v_info->oem_table_id,
> ACPI_OEM_TABLE_ID_SIZE)
> && intel_pstate_no_acpi_pss())
> return true;
> }
>
> if (!strncmp(hdr.oem_id, v_info[1]->oem_id, ACPI_OEM_ID_SIZE) &&
> intel_pstate_has_acpi_ppc())

I really don't think you want to hard code a 1 there.

I think you need to do what Dirk suggested, which is to expand the
hw_vendor_info structure to specify the check that needs to be done
for each entry. For a ProLiant, it would be to call intel_pstate_no_acpi_pss()
and for an Oracle box, it would be to call intel_pstate_has_acpi_ppc().

-- ljk

> return true;
>
> return false;
> }
>
>>>> What about extending the hw_vendor_info struct to include whether _PSS or
>>> Except refer to ACPI DSDT, I don't know how to fill such info.
>>>> _PPC should be done for the platform since it appears that oracle and HP
>>>> have implemented similar functionality using two different methods.
>>> Maybe Linda could answer this whether HP also has _PPC and should be
>>> wept out.
>>> But that doesn't happen with on the same box at the same time.
>> I don't know how an Oracle box works but on a ProLiant, customers can
>> choose to have platform power management or OS power management.
>> When the platform is managing the power, we don't provide the _PSS
>> information. Since our oem information is in the table and there is
>> no _PSS, the intel_pstate driver doesn't stay loaded. That's what we want.
>>
>> When the platform configured to have the OS do the power management,
>> the firmware has _PSS and _PPC and we want the intel_pstate driver,
>> That's what your patch breaks. With your patch, the driver won't
>> stay loaded because our platform is in the table and the check for
>> _PPC passes.
>>
>> How does an Oracle box work?
> Oracle Sun servers (X86) don't have the option in BIOS to change the PM mode
> to firmware/OS,
> The BIOS always has _PSS and _PPC exported to OS whatever 'soft power capping'
> or 'hard power capping' enabled
> in SP configuration web page. if the power policy violation happened, firmware
> will notify OS via SCI with the changed _PPC
> number.
>
> Thanks,
> Ethan
>>
>> -- ljk
>>
>>> Thanks,
>>> Ethan
>>>>
>>>>> I don't suppose you tested on a ProLiant too?
>>>>>
>>>>> -- ljk
>>>>>
>>>>>> + return true;
>>>>>> }
>>>>>>
>>>>>> return false;
>>>>>> }
>>>>>> #else /* CONFIG_ACPI not enabled */
>>>>>> static inline bool intel_pstate_platform_pwr_mgmt_exists(void) {
>>>>>> return false; }
>>>>>> +static inline bool intel_pstate_has_acpi_ppc(void) { return false; }
>>>>>> #endif /* CONFIG_ACPI */
>>>>>>
>>>>>> static int __init intel_pstate_init(void)
>>>>>>
>>> --
>>> To unsubscribe from this list: send the line "unsubscribe linux-pm" in
>>> the body of a message to majordomo@xxxxxxxxxxxxxxx
>>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/