Severe responsiveness issues: irqs disabled after kernel update

[Roger]

Potential Workarounds:

Please try the following two options at your grub command line if you hit this issue:
    pcie_aspm=off
    pcie_aspm=force

If you are affected by this issue please file a new bug for your machine and add the number to this bug for tracking.

===

After todays (27 March) update of the 3.2.0-20-generic x86_64 kernel in Precise, I've started getting messages about irqs (16 and 24 when I had a look after the latest incident) : "nobody cared (try booting with the "irqpoll" option)."

Irq 16 is:

 *-pci:5
             description: PCI bridge
             product: 82801JI (ICH10 Family) PCI Express Root Port 5
             vendor: Intel Corporation
             physical id: 1c.4
             bus info: pci@0000:00:1c.4
             version: 00
             width: 32 bits
             clock: 33MHz
             capabilities: pci pciexpress msi pm normal_decode bus_master cap_list
             configuration: driver=pcieport
             resources: irq:67 ioport:c000(size=4096) memory:fbc00000-fbcfffff ioport:c0200000(size=2097152)
           *-ide
                description: IDE interface
                product: 88SE6121 SATA II Controller
                vendor: Marvell Technology Group Ltd.
                physical id: 0
                bus info: pci@0000:05:00.0
                version: b2
                width: 32 bits
                clock: 33MHz
                capabilities: ide pm msi pciexpress bus_master cap_list
                configuration: driver=pata_marvell latency=0
                resources: irq:16 ioport:cc00(size=8) ioport:c880(size=4) ioport

And irq 24 is:

*-pci:1
             description: PCI bridge
             product: 5520/5500/X58 I/O Hub PCI Express Root Port 3
             vendor: Intel Corporation
             physical id: 3
             bus info: pci@0000:00:03.0
             version: 12
             width: 32 bits
             clock: 33MHz
             capabilities: pci msi pciexpress pm normal_decode bus_master cap_list
             configuration: driver=pcieport
             resources: irq:0 ioport:a000(size=4096) memory:f8000000-fbafffff ioport:d0000000(size=268435456)
           *-display
                description: VGA compatible controller
                product: G92 [GeForce 9800 GT]
                vendor: NVIDIA Corporation
                physical id: 0
                bus info: pci@0000:02:00.0
                version: a2
                width: 64 bits
                clock: 33MHz
                capabilities: pm msi pciexpress vga_controller bus_master cap_list rom
                configuration: driver=nvidia latency=0
                resources: irq:24 memory:fa000000-faffffff memory:d0000000-dffff

The system becomes unresponsive and playing playing a video, for example, becomes impossible since the graphics card is affected. All I get are a still images.

This never happened with any of the previous kernels I've tried in this development series, so it has to be this latest patch that isn't up to par.

The motherboard is an Asus P6T DELUXE V2 with Intel i7 processors.

ProblemType: Bug
DistroRelease: Ubuntu 12.04
Package: linux-image-3.2.0-20-generic 3.2.0-20.33
ProcVersionSignature: Ubuntu 3.2.0-20.33-generic 3.2.12
Uname: Linux 3.2.0-20-generic x86_64
NonfreeKernelModules: nvidia
AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.24.
ApportVersion: 1.95-0ubuntu1
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: roger 2327 F.... pulseaudio
CRDA: Error: [Errno 2] No such file or directory
Date: Wed Mar 28 04:52:18 2012
HibernationDevice: RESUME=UUID=7c7a9b11-96cf-43ee-92e3-1b94656f577e
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Alpha amd64 (20120314)
MachineType: System manufacturer System Product Name
ProcEnviron:
 LANGUAGE=en_GB:en
 TERM=xterm
 LANG=en_GB.UTF-8
 SHELL=/bin/bash
ProcFB: 0 VESA VGA
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-3.2.0-20-generic root=UUID=1bf48e84-7266-4437-a04e-f154423264b5 ro nomodeset quiet splash vt.handoff=7
RfKill:
 0: hci0: Bluetooth
  Soft blocked: no
  Hard blocked: no
SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 03/08/2010
dmi.bios.vendor: American Megatrends Inc.
dmi.bios.version: 1003
dmi.board.asset.tag: To Be Filled By O.E.M.
dmi.board.name: P6T DELUXE V2
dmi.board.vendor: ASUSTeK Computer INC.
dmi.board.version: Rev 1.xx
dmi.chassis.asset.tag: Asset-1234567890
dmi.chassis.type: 3
dmi.chassis.vendor: Chassis Manufacture
dmi.chassis.version: Chassis Version
dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvr1003:bd03/08/2010:svnSystemmanufacturer:pnSystemProductName:pvrSystemVersion:rvnASUSTeKComputerINC.:rnP6TDELUXEV2:rvrRev1.xx:cvnChassisManufacture:ct3:cvrChassisVersion:
dmi.product.name: System Product Name
dmi.product.version: System Version
dmi.sys.vendor: System manufacturer

[Roger (r-wiberg)]

Sorry, I forgot to mention explicitly that the kernel then goes on to disable said irqs:

Disabling IRQ #24

and

Disabling IRQ #16.

[Tim Gardner (timg-tpi)]

The only hardware change from Ubuntu-3.2.0-19.31 to Ubuntu-3.2.0-20.33 that might have an effect is '(pre-stable) ASPM: Fix pcie devices with non-pcie children'. Can you experiment by adding one of these to your grub boot command to see if it makes a difference ?

pcie_aspm=off
pcie_aspm=force

[Roger (r-wiberg)]

@ Tim's message #3: I tried the mainline kernel 3.3.0 for a while as a work-around (worked fine), but went back to 3.2.20-20 to test the boot parameters. So far it seems that both pcie_aspm=force and pcie_aspm=off work fine. Will chime in again if I manage to reproduce the problems with either of the settings.

[Tim Gardner (timg-tpi)]

Roger - it makes sense that 3.3 worked for you since the patch that is causing your regression has not been merged into Linus' tree, though I expect it will soon. On the other hand, this patch fixes boot time crashes for a number of other platforms. I think we're just going to end up release noting the behavior you're seeing. You can add 'pcie_aspm=force' to /etc/default/grub GRUB_CMDLINE_LINUX_DEFAULT to make it permanent. Don't forget to run 'update-grub' afterwards.

[Tim Gardner (timg-tpi)]

This bug should be fixed by "(pre-stable) ASPM: Fix pcie devices with non-pcie children" and pcie_aspm=force in Ubuntu-3.2.0-20.33