Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.
Bug 1345964
Summary: | qemu / kernel hangs when loading ata_piix.ko kernel module when using nested KVM | ||
---|---|---|---|
Product: | Red Hat Enterprise Linux 7 | Reporter: | Matt Flusche <mflusche> |
Component: | qemu-kvm-rhev | Assignee: | Hai Huang <hhuang> |
Status: | CLOSED WONTFIX | QA Contact: | Virtualization Bugs <virt-bugs> |
Severity: | medium | Docs Contact: | |
Priority: | unspecified | ||
Version: | 7.2 | CC: | ablum, bdas, knoel, mflusche, pbonzini, ptoscano, rbalakri, rjones, virt-maint |
Target Milestone: | rc | ||
Target Release: | 7.2 | ||
Hardware: | x86_64 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | If docs needed, set a value | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2016-06-14 07:28:06 UTC | Type: | Bug |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Bug Depends On: | |||
Bug Blocks: | 910269 |
Description
Matt Flusche
2016-06-13 14:04:27 UTC
This is definitely not supposed to happen. It looks like a kernel or qemu bug, with either the kernel or qemu hanging during boot. Probably the simplest thing to check is if there is a new kernel available. Just installing a new kernel and rerunning libguestfs-test-tool will tell you if it's a known kernel bug which has been fixed already. Second thing would be to check if there is a newer qemu, install that and repeat the libguestfs-test-tool command. You can also send me log-in details for the machine where this happens so I can take a closer look (either by email or by private comment on this BZ). Another thing to check is if you have enough free space on /tmp, /var/tmp and/or $TMPDIR if set. qemu hangs if it runs out of space, rather than giving an error. I've tested again with the latest Kernel and qemu-kvm (same results): kernel-3.10.0-327.18.2.el7.x86_64 qemu-kvm-rhev-2.3.0-31.el7_2.13.x86_64 It doesn't seem to be a disk space issue: df -h Filesystem Size Used Avail Use% Mounted on /dev/sda2 25G 5.9G 20G 24% / devtmpfs 7.8G 0 7.8G 0% /dev tmpfs 7.8G 0 7.8G 0% /dev/shm tmpfs 7.8G 384K 7.8G 1% /run tmpfs 7.8G 0 7.8G 0% /sys/fs/cgroup tmpfs 1.6G 0 1.6G 0% /run/user/1000 I'll follow up with access to this env. Thanks for your help. qemu-kvm-rhev-2.3.0-31.el7_2.13.x86_64 This kernel & qemu combo hang when loading the ata_piix.ko module. Note this is under TCG, I have not tested KVM yet. The last messages before the hang are: [ 3.244850] input: PC Speaker as /devices/platform/pcspkr/input/input1 [ 3.244850] Error: Driver 'pcspkr' is already registered, aborting... [ 3.244850] scsi host1: ata_piix [ 3.244850] sd 0:0:0:0: Attached scsi generic sg0 type 0 [ 3.244850] sd 0:0:1:0: Attached scsi generic sg1 type 0 [ 3.244850] scsi host2: ata_piix [ 3.244850] ata1: PATA max MWDMA2 cmd 0x1f0 ctl 0x3f6 bmdma 0xc0a0 irq 14 [ 3.244850] ata2: PATA max MWDMA2 cmd 0x170 ctl 0x376 bmdma 0xc0a8 irq 15 Fails with - qemu-kvm-rhev-2.3.0-31.el7_2.7.x86_64 - qemu-kvm-rhev-2.3.0-31.el7_2.13.x86_64 - kernel-3.10.0-327.10.1.el7.x86_64 - kernel-3.10.0-327.18.2.el7.x86_64 Also fails with - qemu-kvm-rhev-2.1.2-23.el7_1.12.x86_64 I tried to get a stack trace using gdb, but it didn't seem to make sense. With full debugging: [ 0.600000] calling piix_init+0x0/0x1000 [ata_piix] @ 1 [ 0.600000] ata_piix 0000:00:01.1: version 2.13 [ 0.600000] scsi host0: ata_piix [ 0.600000] scsi host1: ata_piix [ 0.600000] ata1: PATA max MWDMA2 cmd 0x1f0 ctl 0x3f6 bmdma 0xc0a0 irq 14 [ 0.600000] ata2: PATA max MWDMA2 cmd 0x170 ctl 0x376 bmdma 0xc0a8 irq 15 [ 0.600000] initcall piix_init+0x0/0x1000 [ata_piix] returned 0 after 50474 usecs <-- hangs here OK, I believe this is a nested KVM issue. I was able to make the test succeed by forcing TCG: LIBGUESTFS_BACKEND_SETTINGS=force_tcg libguestfs-test-tool Matt, could you confirm that nested KVM is being used? You will need to log into the KVM hypervisor host and see if the nested=1 flag is set on the kvm_intel module -- look at the contents of /sys/module/kvm_intel/parameters/nested on the host. I'm afraid that since it seems to be nested KVM, this scenario is both not supported and known to be very troublesome. The only solution is to disable nested KVM. Yes this is nested kvm. This is a training environment and unfortunately requires virtual environments. (In reply to Matt Flusche from comment #11) > Yes this is nested kvm. This is a training environment and unfortunately > requires virtual environments. Sure, but you don't need nested KVM to have nested virtual environments. Turn that setting off and see if it fixes things. Correct; dropping the kvm module; things work better. I believe we will be able to proceed with this configuration. Thanks for your help. # rmmod kvm_amd # time libguestfs-test-tool ************************************************************ * IMPORTANT NOTICE * * When reporting bugs, include the COMPLETE, UNEDITED * output below in your bug report. * ************************************************************ ... ===== TEST FINISHED OK ===== real 0m32.259s user 0m0.130s sys 0m0.472s Closing WONTFIX per comment 9 - not a supported configuration. |