Bug 6716

Summary:	ASPM gets disabled by kernel
Product:	[Hardware Platforms] MinnowBoard MAX Firmware	Reporter:	Marcel Holtmann <marcel>
Component:	minnowboard-uefi-firmware	Assignee:	Darren Hart <dvhart>
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	Medium	CC:	danders, david.wei, dvhart, sjolley.yp.pm, warthog9
Version:	2C A1
Target Milestone:	Production Release
Hardware:	MinnowBoard Max
OS:	Multiple
Whiteboard:
OS type for building Yocto:	---	Type of Regression:	---
Verified:		Documentation change:	No (bug/feature does not impact docs)

Description Marcel Holtmann 2014-09-12 17:38:52 UTC

It seems that ASPM gets disabled due to an invalid UUID:

[    0.734472] acpi PNP0A08:00: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI]
[    0.735321] \_SB_.PCI0:_OSC invalid UUID
[    0.735355] _OSC request data:1 1f 0 
[    0.735378] acpi PNP0A08:00: _OSC failed (AE_ERROR); disabling ASPM

This also has some later impact with the Ethernet controller:

[   13.399731] r8169 Gigabit Ethernet driver 2.3LK-NAPI loaded
[   13.399768] r8169 0000:02:00.0: can't disable ASPM; OS doesn't have ASPM control

Is ASPM suppose to work with these boards?

Comment 1 Darren Hart 2014-09-12 20:21:02 UTC

I hadn't made the connection to the Ethernet driver before this, thanks for that. We had been experiencing some issues with the Ethernet driver during resume that appeared to be resolved in a later kernel, but perhaps this impacted it.

I need to dig into ASPM to know if this is supported. I know we support S3 suspend/resume on this SoC, I'm just not sure exactly what role ASPM plays.

Comment 2 Darren Hart 2014-09-12 20:25:28 UTC

This was also reported here:
https://lists.debian.org/debian-kernel/2014/04/msg00022.html

Some initial research confirms ASPM is indeed supported on the SoC. Leaving this assigned to Tim to investigate the _OSC invalid UUID error.

Comment 3 Darren Hart 2014-11-06 23:24:31 UTC

I no longer see these errors with Linux 3.17.1 nor 3.18-rc3.

I have not seen a statement about a fix in the firmware, it's possible this was a kernel fix.

Tim, has anything changed in the firmware?

I'm prepared to close this as resolved, but would like to here back from Tim first. Assigning to Tim and marking NEEDINFO.

Comment 4 David_Wei 2014-11-07 02:19:53 UTC

Hi Darren,

0.74 binary release have updated _OSC to properly represent support to ASPM and other native PCIe capability. Sorry that we forget to update you in time.

I will change to owner to you on behalf of Tim so that you can review it and then close it.

Comment 5 Darren Hart 2014-11-07 02:25:09 UTC

Thanks David, I verified this afternoon. Marking as closed.