New: Kernelpanic since 3.9.6 with qemu-kvm and pci-passthrough
https://bugzilla.kernel.org/show_bug.cgi?id=60271
Summary: Kernelpanic since 3.9.6 with qemu-kvm and
pci-passthrough
Product: Virtualization
Version: unspecified
Kernel Version: 3.9.8
Platform: All
OS/Version: Linux
Tree: Mainline
Status: NEW
Severity: normal
Priority: P1
Component: kvm
AssignedTo: [email protected]
ReportedBy: [email protected]
Regression: No
I'm (and at least one other user -
https://bugs.archlinux.org/index.php?do=details&task_id=35980) getting kernelpanics since 3.9.8 if I start my VM which is using pci-passthrough. This is new in 3.9.8 - I already can confirm 3.9.6 and 3.9.7 are *not* affected.
Steps to reproduce:
* Boot into kernel 3.9.8
* Start VM - pci-passthrough enabled
* Wait till device is first accessed/used
* Panic/Freeze
I would be glad to help debugging this. Just tell me, which information you need.
Signature · 3 lines
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.6 with qemu-kvm and pci-passthrough
https://bugzilla.kernel.org/show_bug.cgi?id=60271
--- Comment #1 from Fabian Zimmermann <[email protected]> 2013-07-01 17:01:57 --- Created an attachment (id=106521)
--> (https://bugzilla.kernel.org/attachment.cgi?id=106521)
screenshot of panic
Signature · 3 lines
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.6 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
--- Comment #2 from Fabian Zimmermann <[email protected]> 2013-07-01 17:02:23 --- Created an attachment (id=106531)
--> (https://bugzilla.kernel.org/attachment.cgi?id=106531)
dmesg of affected system
Signature · 3 lines
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.6 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
Gleb <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]
--- Comment #3 from Gleb <[email protected]> 2013-07-01 17:33:36 --- You need to catch entire crash message with netconsole or by other means.
Signature · 3 lines
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.6 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
Alex Williamson <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]
--- Comment #4 from Alex Williamson <[email protected]> 2013-07-01 17:50:29 --- There are only two KVM patches between 3.9.7 and 3.9.8, perhaps you could try reverting each of them:
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/patch/?id=c78078398d9c86a7b36b02726399d07013b1e732
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/patch/?id=65f918123bd108a9b50bceaf4b32ac173b9e3f3e
Save to a $FILE, patch -p1 -R < $FILE. I don't see any IOMMU or KVM device assignment specific changes between those stable versions (there are a couple radeon patches too and the archlinux report seems to point at using radeon as an assigned device). As Gleb says, we're just stabbing in the dark w/o seeing the beginning of panic message.
Signature · 3 lines
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.6 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
--- Comment #5 from Fabian Zimmermann <[email protected]> 2013-07-01 18:56:45 --- tried to get dmesg with netconsole, but didn't got any "verbose" msg in dmesg/console, only:
[ 214.717070] input: PC Speaker as /devices/platform/pcspkr/input/input7 [ 220.578854] input: PC Speaker as /devices/platform/pcspkr/input/input8 [ 222.689153] input: PC Speaker as /devices/platform/pcspkr/input/input9 [ 237.756584] cgroup: libvirtd (797) created nested cgroup for controller "memory" which has incomplete hierarchy support. Nested cgroups may change behavior in the future. [ 237.756595] cgroup: "memory" requires setting use_hierarchy to 1 on the root. [ 237.756651] cgroup: libvirtd (797) created nested cgroup for controller "devices" which has incomplete hierarchy support. Nested cgroups may change behavior in the future. [ 237.756688] cgroup: libvirtd (797) created nested cgroup for controller "blkio" which has incomplete hierarchy support. Nested cgroups may change behavior in the future. [ 237.758818] tun: Universal TUN/TAP device driver, 1.6 [ 237.758825] tun: (C) 1999-2004 Max Krasnyansky <[email protected]> [ 237.778667] device vnet0 entered promiscuous mode [ 237.801933] br0: port 2(vnet0) entered forwarding state [ 237.801946] br0: port 2(vnet0) entered forwarding state [ 237.828833] qemu-system-x86: sending ioctl 5326 to a partition! [ 237.828851] qemu-system-x86: sending ioctl 80200204 to a partition! [ 237.949768] pci-stub 0000:01:00.0: enabling device (0000 -> 0003) [ 238.586632] assign device 0:1:0.0 [ 238.734338] assign device 0:1:0.1 [ 238.837611] assign device 0:0:1b.0 [ 238.940852] assign device 0:0:1a.0 [ 252.827533] br0: port 2(vnet0) entered forwarding state [ 267.843359] pci-stub 0000:01:00.0: irq 47 for MSI/MSI-X
... general protection fault: 0000 [#1]
will now try to reverse above patches.
Signature · 3 lines
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.8 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
Fabian Zimmermann <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|Kernelpanic since 3.9.6 |Kernelpanic since 3.9.8
|with qemu-kvm and |with qemu-kvm and
|pci-passthrough |pci-passthroughSignature · 2 lines
You are receiving this mail because: You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.8 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
--- Comment #6 from Fabian Zimmermann <[email protected]> ---
* reversed both above patches - problem still there
* disabled radeon-module (in kernel-config) - problem still there
Attached you will find the dmesg.txt (netconsole-output of panic). Don't hesitate to ask if I can provide further information
Signature · 2 lines
You are receiving this mail because: You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.8 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
--- Comment #7 from Fabian Zimmermann <[email protected]> --- Created attachment 106838
--> https://bugzilla.kernel.org/attachment.cgi?id=106838&action=edit
netconsole / dmesg of panic
Signature · 2 lines
You are receiving this mail because: You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.8 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
Fabian Zimmermann <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
Kernel Version|3.9.8 |3.9.8, 3.9.9
--- Comment #8 from Fabian Zimmermann <[email protected]> --- 3.9.9 is affected, too.
Signature · 2 lines
You are receiving this mail because: You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.8 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
Michael S. Tsirkin <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]
--- Comment #9 from Michael S. Tsirkin <[email protected]> --- can you pls try disabling zero copy tx in vhost_net? it's a module parameter for this module
Signature · 2 lines
You are receiving this mail because: You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.8 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
--- Comment #10 from Fabian Zimmermann <[email protected]> --- adding "vhost_net.experimental_zcopytx=0" to kernel-cmdline solved the issue!
Anything else I may do/test?
Signature · 2 lines
You are receiving this mail because: You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.8 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
--- Comment #11 from Michael S. Tsirkin <[email protected]> --- could you test latest upstream please?
A fix from there is in the process of being backported to stable.
Signature · 2 lines
You are receiving this mail because: You are watching the assignee of the bug.
[Bug 60271] Kernelpanic since 3.9.8 with qemu-kvm and pci-passthrough
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=60271
Fabian Zimmermann <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution|--- |PATCH_ALREADY_AVAILABLE--- Comment #12 from Fabian Zimmermann <[email protected]> --- Bug is gone in 3.11-rc1.
Thanks a lot!
Signature · 2 lines
You are receiving this mail because: You are watching the assignee of the bug.