hard lock kernel

[swmike]

Ubuntu 16.04 guest VM running under vmware hypervisor. VM is used for running qemu/docker instances of virtual routers.

After upgrade to linux-image-4.4.0-184-generic we get hard-lockup of the kernel (no interaction on the console, no crash report or anything) as soon as VM is put under load (several minutes in, 5 virtual routers started and configured).

VM was running -177 or -178 for ~50 days without issue. I tested -179 as well, doesn't have this problem. linux-image-4.15.0-106-generic (latest hwe kernel) also doesn't have this problem. All of these are Ubuntu provided kernels installed using regular install process.

So to sum up:

Works: 4.4.0-177 to -179. 4.15.0-106 also works.

Doesn't work: 4.4.0-184 causes immediate problems (within 10 minutes after starting this test/simulation environment). Hard-lockup only happens during the test and seems to be either memory/load dependent.

Unfortunately I don't have any debug output because of this hard-lockup.

$ sudo dpkg --list | grep -i linux-image
ii linux-image-4.15.0-106-generic 4.15.0-106.107~16.04.1 amd64
Signed kernel image generic
rc linux-image-4.4.0-177-generic 4.4.0-177.207 amd64 Signed kernel image generic
rc linux-image-4.4.0-178-generic 4.4.0-178.208 amd64 Signed kernel image generic
ii linux-image-4.4.0-179-generic 4.4.0-179.209 amd64 Signed kernel image generic
ii linux-image-4.4.0-184-generic 4.4.0-184.214 amd64 Signed kernel image generic
ii linux-image-generic 4.4.0.184.190 amd64 Generic Linux kernel image
ii linux-image-generic-hwe-16.04 4.15.0.106.111 amd64 Generic Linux kernel image

ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: linux-image-4.4.0-184-generic 4.4.0-184.214
ProcVersionSignature: Ubuntu 4.15.0-106.107~16.04.1-generic 4.15.18
Uname: Linux 4.15.0-106-generic x86_64
ApportVersion: 2.20.1-0ubuntu2.23
Architecture: amd64
Date: Fri Jun 26 07:10:53 2020
ProcEnviron:
 TERM=xterm-256color
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: linux-signed
UpgradeStatus: No upgrade log present (probably fresh install)