Without this patchset applied to the L1 VM kernel L2 VMs may crash on boot in case L1 VM is reset using following command: # echo b > /proc/sysrq-trigger
Following message appeared in the L2 VM log: KVM: entry failed, hardware error 0x80000021 https://pmc.acronis.work/browse/VSTOR-76102 Note: all patches except "(x86/reboot: VMCLEAR active VMCSes before emergency reboot") are present in the latest CentOS 9 Stream kernel 5.14.0-381.el9, so they are to be dropped during future rebases vz9 kernel onto new RHEL9 kernels for sure. The issue had been reported to RedHat as well: https://issues.redhat.com/browse/RHEL-14437 Maxim Levitsky (8): ms/x86/crash: Disable virt in core NMI crash handler to avoid double shootdown ms/x86/virt: Force GIF=1 prior to disabling SVM (for reboot flows) ms/x86/reboot: Disable virtualization in an emergency if SVM is supported ms/x86/reboot: Disable SVM, not just VMX, when stopping CPUs ms/KVM: VMX: Reset eVMCS controls in VP assist page during hardware disabling ms/KVM: VMX: Don't bother disabling eVMCS static key on module exit ms/KVM: x86: Move guts of kvm_arch_init() to standalone helper ms/KVM: VMX: Do _all_ initialization before exposing /dev/kvm to userspace Sean Christopherson (1): ms/x86/reboot: VMCLEAR active VMCSes before emergency reboot arch/x86/include/asm/kexec.h | 2 - arch/x86/include/asm/kvm_host.h | 3 + arch/x86/include/asm/reboot.h | 4 ++ arch/x86/include/asm/virtext.h | 16 ++++- arch/x86/kernel/crash.c | 48 +------------- arch/x86/kernel/reboot.c | 110 +++++++++++++++++++++++++------- arch/x86/kernel/smp.c | 6 +- arch/x86/kvm/svm/svm.c | 23 ++++++- arch/x86/kvm/vmx/vmx.c | 101 +++++++++++++++++------------ arch/x86/kvm/x86.c | 15 ++++- 10 files changed, 206 insertions(+), 122 deletions(-) -- 2.39.3 _______________________________________________ Devel mailing list Devel@openvz.org https://lists.openvz.org/mailman/listinfo/devel