Bug 69345 - nouveau no longer reports D3cold on linux-next as of at least 20130906
Summary: nouveau no longer reports D3cold on linux-next as of at least 20130906
Status: RESOLVED NOTABUG
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/nouveau (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium minor
Assignee: Nouveau Project
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-09-14 06:38 UTC by Tod Jackson
Modified: 2013-10-11 00:57 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
current dmesg (83.84 KB, text/plain)
2013-09-14 06:56 UTC, Tod Jackson
Details

Description Tod Jackson 2013-09-14 06:38:57 UTC
Hi there,

I saw Rafael made this recent commit: http://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/commit/?id=8ad928d52e63a9b7d69f0873d7318c4561e2f8cd

That may not be the culprit at all, sorry. I noticed that if I'm using, say, the latest drm-intel-next kernel or nouveau kernel ACPI reports nouveau's status properly, e.g. nouveau: power state changed by ACPI to D3cold, however in next I no longer get these messages, although the device seems to be in D3cold (?). At least temperatures are much lower, especially when building software.

[23301.994315] ACPI Warning: \_SB_.PCI0.P0P2.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130725/nsarguments-95)
(That's the only indicator of power transitions now - a warning that might go away in the future - p.s., do I need to do some DSDT hacking for that or is it on the Linux-side?)

Invoking lm_sensors' sensors command shows N/A for the nouveau device on Optimus laptops, as expected if the device is in D3cold.
Comment 1 Tod Jackson 2013-09-14 06:56:10 UTC
Created attachment 85810 [details]
current dmesg
Comment 2 Peter Wu 2013-09-16 10:41:08 UTC
That DSM argument mismatch warning is caused by the modules (nouveau and bbswitch) and is a workaround for buggy DSDTs. It is possible to make the warning go away, but that will confuse machines with buggy DSDTs.

I haven't looked in the code yet, but what is the issue with D3cold not being mentioned in dmesg?
Comment 3 Tod Jackson 2013-09-16 15:54:41 UTC
It's quite possibly a non-issue. But it's new behavior, so I just wanted to check it was intentional.
Comment 4 Ilia Mirkin 2013-10-11 00:57:47 UTC
I believe this is part of the runtime pm support that was added for Optimus setups in 3.12-rc1, which puts the nvidia card into d3cold when it's not being used. You can disable it by adding nouveau.runpm=1. If there are any actual issues with it, feel free to open issues about it. I think that ACPI warning is about buggy AML, perhaps a BIOS update will fix it, or more likely, you'll have to live with it.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.