It was reported that when a guest kernel running under hypervisor-mode KVM (i.e. "HV KVM" rather than "PR KVM"), does the H_CEDE hypercall while a suspended transaction exists, the host CPU may, under certain circumstances, hang in an infinite loop with interrupts disabled. To trigger this, it is necessary to have the guest using an SMT mode of at least 2, with another thread in the same core busy at the time that the H_CEDE is done. The other thread doesn't need to be using transactional memory.
Created attachment 1171980 [details]
Created attachment 1171981 [details]
Created kernel tracking bugs for this issue:
Affects: fedora-all [bug 1361040]
This issue does not affect the versions of Linux kernel as shipped with
Red Hat Enterprise Linux 5, 6 and Red Hat Enterprise MRG 2.
This issue affects the version of the kernel packages as shipped with
Red Hat Enterprise Linux 7. Future kernel updates for Red Hat Enterprise Linux 7 may address this issue.
kernel-4.6.6-300.fc24 has been pushed to the Fedora 24 stable repository. If problems still persist, please make note of it in this bug report.
kernel-4.6.6-200.fc23 has been pushed to the Fedora 23 stable repository. If problems still persist, please make note of it in this bug report.
This issue has been addressed in the following products:
Red Hat Enterprise Linux 7
Via RHSA-2016:2574 https://rhn.redhat.com/errata/RHSA-2016-2574.html