USB power not turned off using nForce 590 SLI chipset

[Ian McMichael]

I have a Gigabyte GA-M59SLI-S5 motherboard with BIOS F9c upon which I am now running Natty (11.04). Devices like the keyboard (Microsoft Comfort Curve) and mouse (Microsoft Wheel Mouse Optical) connected to the USB bus are no longer powered off at shutdown.

No hardware configuration was changed from Maverick (10.10) to Natty (11.04) and this issue now occurs. To confirm the options in the BIOS to allow resume from mouse and/or keyboard are both disabled, which can also display this behaviour. For this reason I believe this to be a regression somewhere in the power control of the kernel.

ProblemType: Bug
DistroRelease: Ubuntu 11.04
Package: linux-image-2.6.38-8-generic 2.6.38-8.42
ProcVersionSignature: Ubuntu 2.6.38-8.42-generic 2.6.38.2
Uname: Linux 2.6.38-8-generic i686
NonfreeKernelModules: nvidia
AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.23.
Architecture: i386
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: ian 1849 F.... pulseaudio
 /dev/snd/controlC1: ian 1849 F.... pulseaudio
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
 Card hw:0 'NVidia'/'HDA NVidia at 0xfd100000 irq 20'
   Mixer name : 'Realtek ALC888'
   Components : 'HDA:10ec0888,1458e601,00100001'
   Controls : 40
   Simple ctrls : 22
Card1.Amixer.info:
 Card hw:1 'U0x46d0x807'/'USB Device 0x46d:0x807 at usb-0000:00:0a.1-7, high speed'
   Mixer name : 'USB Mixer'
   Components : 'USB046d:0807'
   Controls : 2
   Simple ctrls : 1
Card1.Amixer.values:
 Simple mixer control 'Mic',0
   Capabilities: cvolume cvolume-joined cswitch cswitch-joined penum
   Capture channels: Mono
   Limits: Capture 0 - 6144
   Mono: Capture 6144 [100%] [30.00dB] [on]
Date: Tue May 24 15:30:04 2011
HibernationDevice: RESUME=UUID=8bca4d7b-e630-4aee-85c5-697fc2187ab2
InstallationMedia: Ubuntu 11.04 "Natty Narwhal" - Release i386 (20110427.1)
ProcEnviron:
 LANGUAGE=en_GB:en
 LANG=en_GB.UTF-8
 SHELL=/bin/bash
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.38-8-generic root=UUID=80dec9f0-ab97-403a-8cca-237f5b2e75a4 ro quiet splash vt.handoff=7
RelatedPackageVersions:
 linux-restricted-modules-2.6.38-8-generic N/A
 linux-backports-modules-2.6.38-8-generic N/A
 linux-firmware 1.52
RfKill:

SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 06/18/2008
dmi.bios.vendor: Award Software International, Inc.
dmi.bios.version: F9c
dmi.board.name: M59SLI-S5
dmi.board.vendor: Gigabyte Technology Co., Ltd.
dmi.chassis.type: 3
dmi.modalias: dmi:bvnAwardSoftwareInternational,Inc.:bvrF9c:bd06/18/2008:svn:pn:pvr:rvnGigabyteTechnologyCo.,Ltd.:rnM59SLI-S5:rvr:cvn:ct3:cvr:

[Ian McMichael (ian-sigma-uk)]

Having dug a little deeper I have found the following kernel patch:

http://www.spinics.net/lists/linux-usb/msg35958.html

Can anyone walk me through finding out if it was included in 2.6 kernel releases up until Natty? If so, I will then need to work out why it was removed for Natty and if it can safely be requested to be put back.

[Julian Wiedmann (jwiedmann)]

Natty actually is the *first* release with that patch included (although it has propagated to lucid, through -stable), fixing the issues you described for older Nvidia controllers. The (more recent) USB chip on your system doesn't require this workaround (that's why previous releases worked just fine) - but as it's currently applied for *all* Nvidia chips, bad things happen on your machine.

For details, see bug 42149 and:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=2b7aaf503d56216b847c8265421d2a7d9b42df3e

That fix-up patch was merged for 2.6.40 and is CCed to stable.

[Ian McMichael (ian-sigma-uk)]

Thanks, that's excellent news and fully explains why I am not suffering difficulties with the Natty kernel. What do I need to do to get that patch included in the Natty kernel to fix this regression?

[Ian McMichael (ian-sigma-uk)]

* not = now in the previous post. Sorry for this confusing typo!

[Julian Wiedmann (jwiedmann)]

It should be picked up for one of the next 2.6.38-stable releases, and eventually make its way into natty from there.

Fast track? Have a member of the kernel team pick this up as pre-stable, so it get's merged into the next -proposed kernel immediately.

[Ian McMichael (ian-sigma-uk)]

Thanks, that's good news. As a humble user (with a fair degree of hardware and development knowledge) I am not on personal terms with any members of the kernel team. I am guessing their group is subscribed to this bug? If so, could anyone assist me in getting this patch moved into pre-stable (whatever that is!)?

[Seth Forshee (sforshee)]

It's been submitted to stable but isn't queued up yet. I'll keep an eye on it and update the status when it makes it into our natty tree.

[Seth Forshee (sforshee)]

I've posted a test build with the fix applied for testing.

http://people.canonical.com/~sforshee/lp787590/linux-2.6.38-9.43~lp787590v201106021536/

Please post your results back here. Thanks!

[Ian McMichael (ian-sigma-uk)]

Sorry for the delay, I have been away from my nForce 590 based machine for the last few days. I have now tested your kernel and it fixes this regression with no side-effects that I could see. USB devices are correctly powered off when the system shuts down.

Does this mean we will see this in an official Natty kernel in the near future?

[Seth Forshee (sforshee)]

The fix is in the latest proposed kernel, which means it will be included in the next kernel release as long as it isn't found to cause any problems. This kernel should be released in the next 2-3 weeks.

[Seth Forshee (sforshee)]

I made a mistake. The fix is not currently in proposed, but will be included in the next proposed kernel (should appear in proposed about 3 weeks from now). Sorry for the confusion!

[Ian McMichael (ian-sigma-uk)]

Thanks for letting me know. Is there a way I can see when it is proposed to test it there or will you update the bug when it gets included?

[Seth Forshee (sforshee)]

@Ian, I don't know of a good way for you to know when it gets to proposed. I'll do my best to post an update here once it gets to proposed.

[Ian McMichael (ian-sigma-uk)]

Sadly today's security update for the kernel (2.6.38-10.46) has knocked out this patch and the problem is back. Is there any chance you could build another image based on this kernel? If not, do you have a diff I can apply to the kernel package and I'll have a go at building one in my PPA?

[Seth Forshee (sforshee)]

Ian: You should still be able to select the other kernel by holding Shift just after the BIOS screen disappears until you get the boot menu, then select "Previous Linux versions", and then select the patched kernel. The fix will be in the next proposed kernel, so it shouldn't be too much longer.

[Ian McMichael (ian-sigma-uk)]

Thanks, that's certainly a work-around. It does mean that the latest set of security patches won't be in the running kernel then. Does the fact that a kernel has just been released mean the new proposed one including this patch will be built in the next day or so?

[Seth Forshee (sforshee)]

I know it's already been uploaded, so I'd expect it to hit proposed
before too much longer.

[Julian Wiedmann (jwiedmann)]

Available in -proposed now.

[Ian McMichael (ian-sigma-uk)]

Thanks Julian. Following the instructions at https://wiki.ubuntu.com/Testing/EnableProposed I have upgraded to kernel 2.6.38-11.47, which correctly shuts the power down to USB devices at shutdown.

[Ian McMichael (ian-sigma-uk)]

Tested and also working correctly on today's 2.6.38-11.48 proposed kernel.

[Julian Wiedmann (jwiedmann)]

2.6.38-11.48 has been copied to -updates.