Re: [RFC] PCIE ASPM support

["Kok, Auke" <auke-jan.h.kok@intel.com>]

Shaohua Li wrote:
> On Thu, 2008-01-03 at 11:33 -0800, Kok, Auke wrote:
>> Shaohua Li wrote:
>>> PCI Express ASPM defines a protocol for PCI Express components in the D0
>>> state to reduce Link power by placing their Links into a low power state
>>> and instructing the other end of the Link to do likewise. This
>>> capability allows hardware-autonomous, dynamic Link power reduction
>>> beyond what is achievable by software-only controlled power management.
>>> However, The device should be configured by software appropriately.
>>> Enabling ASPM will save power, but will introduce device latency.
>>>
>>> This patch adds ASPM support in Linux. It introduces a global policy for
>>> ASPM, a sysfs file /sys/module/pcie_aspm/parameters/policy can control
>>> it. The interface can be used as a boot option too. Currently we have
>>> below setting:
>>> 	-default, BIOS default setting
>>> 	-powersave, highest power saving mode, enable all available ASPM state
>>> and clock power management
>>> 	-performance, highest performance, disable ASPM and clock power
>>> management
>>> By default, the 'default' policy is used currently.
>>>
>>> In my test, power difference between powersave mode and performance mode
>>> is about 1.3w in a system with 3 PCIE links.
>>>
>>> please review, any comments will be appreciated.
>>
>> quickly glanced this over since I recently disabled l1 ASPM for the e1000/e1000e
>> driven 82573 device which has issues with l1 ASPM. that immediately gives me the
>> question: how can I continue to disable 1l aspm by default for this device using
>> this infrastructure?
> I used to have a per-device interface, but thought the interface might
> be hard to use for users. If we really need the per-device interface, I
> can re-add it.
> 
>> I do like the fact that there is a generic way to re-enable it for the users who
>> want to use it. Can this change be done when the device is already active?
> Yes, at least in my test.
> 
>> Can you
>> change this parameter per device/module?
> Another way is to provide a helper for driver, and driver disables
> specific ASPM states. It sounds better to let driver do the disabling,
> as users haven't the knowledge?

agreed, however this could still be usefull in debugging equipment for the
experienced user. In any case an easy handle for the driver to dis/enable ASPM
would certainly help our case, and possibly others.

Auke