Bug 1397296
Summary: | :KVM internal error. Suberror: 1 emulation failure | ||
---|---|---|---|
Product: | Red Hat Enterprise Linux 7 | Reporter: | Jaroslav Kortus <jkortus> |
Component: | seabios | Assignee: | Radim Krčmář <rkrcmar> |
Status: | CLOSED DUPLICATE | QA Contact: | Virtualization Bugs <virt-bugs> |
Severity: | high | Docs Contact: | |
Priority: | high | ||
Version: | 7.3 | CC: | bsd, chayang, dgilbert, emahoney, jkortus, juzhang, knoel, pbonzini, rbalakri, shivrao, virt-maint |
Target Milestone: | rc | ||
Target Release: | --- | ||
Hardware: | Unspecified | ||
OS: | Unspecified | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | If docs needed, set a value | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2017-04-12 13:11:31 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: |
Description
Jaroslav Kortus
2016-11-22 08:17:18 UTC
Can you install a RHEL 7.2 kernel on that 7.3 machine to verify that it is a regression in the kernel? The VM fails after the reboot and before Linux is loaded? Thanks. when I was observing one virt crashing, it crashed in the middle of the boot (right after the serial driver is loaded). Unfortunately it was not crashing consistently and I was not able to come up with some more reliable reproducer. I will give a try to the combination you suggested. Observations and questions... Machine type is rhel6.3.0. <type arch='x86_64' machine='rhel6.3.0'>hvm</type> What is the guest OS version? Memory hard limit is higher than current memory: <currentMemory unit='KiB'>1048576</currentMemory> <hard_limit unit='KiB'>1073741824</hard_limit> 1 vcpu: <vcpu placement='static'>1</vcpu> No cpu model specified. What is the QEMU command line? What is the host processor? Can be reproduced at Red Hat: <name>virt-361.cluster-qe.lab.eng.brq.redhat.com</name> No z-stream packages installed: kernel-3.10.0-514.el7.x86_64 libvirt-2.0.0-10.el7.x86_64 qemu-img-1.5.3-126.el7.x86_64 qemu-kvm-1.5.3-126.el7.x86_64 Can you update to the latest z-stream and retry? Can a KVM engineer borrow the machine to debug? The customer hit emulation error on "fadds" instruction (d8 01), which might point to problems with FPU. Can you reproduce with "noxsave" kernel parameter in host and guest? The internal error doesn't really say what happened, so providing access to a reproducer would be great as there are many things to try. Thanks. The original report indicates that the bug did not happen on RHEL 7.2, so downgrading kernel or qemu-kvm RPMs could serve as a work-around. If that did not help, we would likely have two bugs in this BZ. Preliminary testing with 7.4 seabios went well, so here is a backport of two patches (the first mostly for context, the second one should fix this issue) to 7.3 seabios for more thorough testing: b837e68d5a6c ("resume: Make KVM soft reboot loop detection more flexible") c68aff57ce31 ("resume: Don't attempt to use generic reboot mechanisms on QEMU") Please try the following RPM: http://people.redhat.com/rkrcmar/seabios-bin-1.9.1-5.el7_3.1.bz1397296.noarch.rpm I would like to request a backport to 7.3 for this fix. Shiva, has the custom 7.3 seabios rpm fixed the bug for your setup? It didn't help in the linked support case, so I'm wondering what is different. Thanks. After debugging the setup we see this time the issue was related to the "arat" flag being passed to the guest. So I can confirm we are not hitting this issue anymore. Coudl we please backport this to RHEL 7.3 Thank you. The 7.3 backport is tracked under BZ 1440706 and has already been posted for inclusion in the next 7.3.z. I am closing this as a duplicate of BZ 1428347. *** This bug has been marked as a duplicate of bug 1428347 *** |